ruby_llm 1.4.0 → 1.5.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
@@ -963,9 +963,7 @@
963
963
  },
964
964
  "metadata": {
965
965
  "provider_name": "Anthropic",
966
- "inference_types": [
967
- "PROVISIONED"
968
- ],
966
+ "inference_types": [],
969
967
  "streaming_supported": true,
970
968
  "input_modalities": [
971
969
  "TEXT",
@@ -1014,9 +1012,7 @@
1014
1012
  },
1015
1013
  "metadata": {
1016
1014
  "provider_name": "Anthropic",
1017
- "inference_types": [
1018
- "PROVISIONED"
1019
- ],
1015
+ "inference_types": [],
1020
1016
  "streaming_supported": true,
1021
1017
  "input_modalities": [
1022
1018
  "TEXT",
@@ -1065,9 +1061,7 @@
1065
1061
  },
1066
1062
  "metadata": {
1067
1063
  "provider_name": "Anthropic",
1068
- "inference_types": [
1069
- "PROVISIONED"
1070
- ],
1064
+ "inference_types": [],
1071
1065
  "streaming_supported": true,
1072
1066
  "input_modalities": [
1073
1067
  "TEXT",
@@ -2164,40 +2158,33 @@
2164
2158
  },
2165
2159
  {
2166
2160
  "id": "gemini-1.5-flash-8b-latest",
2167
- "name": "Gemini 1.5 Flash-8B Latest",
2161
+ "name": "Gemini 1.5 Flash-8B",
2168
2162
  "provider": "gemini",
2169
- "family": "gemini15_flash_8b",
2163
+ "family": "gemini-1.5-flash-8b",
2170
2164
  "created_at": null,
2171
- "context_window": 1000000,
2165
+ "context_window": 1048576,
2172
2166
  "max_output_tokens": 8192,
2173
2167
  "knowledge_cutoff": null,
2174
2168
  "modalities": {
2175
2169
  "input": [
2176
- "text",
2170
+ "audio",
2177
2171
  "image",
2178
- "pdf"
2172
+ "text"
2179
2173
  ],
2180
2174
  "output": [
2181
2175
  "text"
2182
2176
  ]
2183
2177
  },
2184
2178
  "capabilities": [
2185
- "streaming",
2186
2179
  "function_calling",
2187
- "structured_output",
2188
- "batch",
2189
- "caching",
2190
- "fine_tuning"
2180
+ "structured_output"
2191
2181
  ],
2192
2182
  "pricing": {
2193
2183
  "text_tokens": {
2194
2184
  "standard": {
2195
- "input_per_million": 0.0375,
2196
- "output_per_million": 0.15
2197
- },
2198
- "batch": {
2199
- "input_per_million": 0.01875,
2200
- "output_per_million": 0.075
2185
+ "input_per_million": 0.075,
2186
+ "cached_input_per_million": 0.01875,
2187
+ "output_per_million": 0.3
2201
2188
  }
2202
2189
  }
2203
2190
  },
@@ -2213,40 +2200,33 @@
2213
2200
  },
2214
2201
  {
2215
2202
  "id": "gemini-1.5-flash-latest",
2216
- "name": "Gemini 1.5 Flash Latest",
2203
+ "name": "Gemini 1.5 Flash",
2217
2204
  "provider": "gemini",
2218
- "family": "gemini15_flash",
2205
+ "family": "gemini-1.5-flash",
2219
2206
  "created_at": null,
2220
- "context_window": 1000000,
2207
+ "context_window": 1048576,
2221
2208
  "max_output_tokens": 8192,
2222
2209
  "knowledge_cutoff": null,
2223
2210
  "modalities": {
2224
2211
  "input": [
2225
- "text",
2212
+ "audio",
2226
2213
  "image",
2227
- "pdf"
2214
+ "text"
2228
2215
  ],
2229
2216
  "output": [
2230
2217
  "text"
2231
2218
  ]
2232
2219
  },
2233
2220
  "capabilities": [
2234
- "streaming",
2235
2221
  "function_calling",
2236
- "structured_output",
2237
- "batch",
2238
- "caching",
2239
- "fine_tuning"
2222
+ "structured_output"
2240
2223
  ],
2241
2224
  "pricing": {
2242
2225
  "text_tokens": {
2243
2226
  "standard": {
2244
2227
  "input_per_million": 0.075,
2228
+ "cached_input_per_million": 0.01875,
2245
2229
  "output_per_million": 0.3
2246
- },
2247
- "batch": {
2248
- "input_per_million": 0.0375,
2249
- "output_per_million": 0.15
2250
2230
  }
2251
2231
  }
2252
2232
  },
@@ -2378,38 +2358,33 @@
2378
2358
  },
2379
2359
  {
2380
2360
  "id": "gemini-1.5-pro-latest",
2381
- "name": "Gemini 1.5 Pro Latest",
2361
+ "name": "Gemini 1.5 Pro",
2382
2362
  "provider": "gemini",
2383
- "family": "gemini15_pro",
2363
+ "family": "gemini-1.5-pro",
2384
2364
  "created_at": null,
2385
- "context_window": 2000000,
2365
+ "context_window": 2097152,
2386
2366
  "max_output_tokens": 8192,
2387
2367
  "knowledge_cutoff": null,
2388
2368
  "modalities": {
2389
2369
  "input": [
2390
- "text",
2370
+ "audio",
2391
2371
  "image",
2392
- "pdf"
2372
+ "text"
2393
2373
  ],
2394
2374
  "output": [
2395
2375
  "text"
2396
2376
  ]
2397
2377
  },
2398
2378
  "capabilities": [
2399
- "streaming",
2400
2379
  "function_calling",
2401
- "structured_output",
2402
- "caching"
2380
+ "structured_output"
2403
2381
  ],
2404
2382
  "pricing": {
2405
2383
  "text_tokens": {
2406
2384
  "standard": {
2407
2385
  "input_per_million": 1.25,
2386
+ "cached_input_per_million": 0.3125,
2408
2387
  "output_per_million": 5.0
2409
- },
2410
- "batch": {
2411
- "input_per_million": 0.625,
2412
- "output_per_million": 2.5
2413
2388
  }
2414
2389
  }
2415
2390
  },
@@ -3626,26 +3601,30 @@
3626
3601
  },
3627
3602
  {
3628
3603
  "id": "gemini-embedding-001",
3629
- "name": "Gemini Embedding",
3604
+ "name": "Gemini Embedding 001",
3630
3605
  "provider": "gemini",
3631
- "family": "gemini-embedding-001",
3606
+ "family": "embedding1",
3632
3607
  "created_at": null,
3633
3608
  "context_window": 2048,
3634
- "max_output_tokens": null,
3609
+ "max_output_tokens": 1,
3635
3610
  "knowledge_cutoff": null,
3636
3611
  "modalities": {
3637
3612
  "input": [
3638
3613
  "text"
3639
3614
  ],
3640
3615
  "output": [
3616
+ "text",
3641
3617
  "embeddings"
3642
3618
  ]
3643
3619
  },
3644
- "capabilities": [],
3620
+ "capabilities": [
3621
+ "streaming",
3622
+ "batch"
3623
+ ],
3645
3624
  "pricing": {
3646
- "text_tokens": {
3625
+ "embeddings": {
3647
3626
  "standard": {
3648
- "input_per_million": 0.15
3627
+ "input_per_million": 0.002
3649
3628
  }
3650
3629
  }
3651
3630
  },
@@ -3715,26 +3694,45 @@
3715
3694
  },
3716
3695
  {
3717
3696
  "id": "gemini-embedding-exp-03-07",
3718
- "name": "Gemini Embedding",
3697
+ "name": "Gemini Embedding Experimental 03-07",
3719
3698
  "provider": "gemini",
3720
- "family": "gemini-embedding-001",
3699
+ "family": "gemini_embedding_exp",
3721
3700
  "created_at": null,
3722
- "context_window": 2048,
3723
- "max_output_tokens": null,
3701
+ "context_window": 8192,
3702
+ "max_output_tokens": 1,
3724
3703
  "knowledge_cutoff": null,
3725
3704
  "modalities": {
3726
3705
  "input": [
3727
- "text"
3706
+ "text",
3707
+ "image",
3708
+ "pdf"
3728
3709
  ],
3729
3710
  "output": [
3711
+ "text",
3730
3712
  "embeddings"
3731
3713
  ]
3732
3714
  },
3733
- "capabilities": [],
3715
+ "capabilities": [
3716
+ "streaming",
3717
+ "function_calling",
3718
+ "structured_output",
3719
+ "batch",
3720
+ "caching"
3721
+ ],
3734
3722
  "pricing": {
3735
3723
  "text_tokens": {
3736
3724
  "standard": {
3737
- "input_per_million": 0.15
3725
+ "input_per_million": 0.002,
3726
+ "output_per_million": 0.004
3727
+ },
3728
+ "batch": {
3729
+ "input_per_million": 0.001,
3730
+ "output_per_million": 0.002
3731
+ }
3732
+ },
3733
+ "embeddings": {
3734
+ "standard": {
3735
+ "input_per_million": 0.002
3738
3736
  }
3739
3737
  }
3740
3738
  },
@@ -4027,317 +4025,2147 @@
4027
4025
  }
4028
4026
  },
4029
4027
  "metadata": {
4030
- "version": "001",
4031
- "description": null,
4032
- "supported_generation_methods": [
4033
- "generateContent",
4034
- "countTokens"
4035
- ]
4028
+ "version": "001",
4029
+ "description": null,
4030
+ "supported_generation_methods": [
4031
+ "generateContent",
4032
+ "countTokens"
4033
+ ]
4034
+ }
4035
+ },
4036
+ {
4037
+ "id": "gemma-3n-e4b-it",
4038
+ "name": "Gemma 3n E4B",
4039
+ "provider": "gemini",
4040
+ "family": "other",
4041
+ "created_at": null,
4042
+ "context_window": 8192,
4043
+ "max_output_tokens": 2048,
4044
+ "knowledge_cutoff": null,
4045
+ "modalities": {
4046
+ "input": [
4047
+ "text"
4048
+ ],
4049
+ "output": [
4050
+ "text"
4051
+ ]
4052
+ },
4053
+ "capabilities": [
4054
+ "streaming"
4055
+ ],
4056
+ "pricing": {
4057
+ "text_tokens": {
4058
+ "standard": {
4059
+ "input_per_million": 0.075,
4060
+ "output_per_million": 0.3
4061
+ },
4062
+ "batch": {
4063
+ "input_per_million": 0.0375,
4064
+ "output_per_million": 0.15
4065
+ }
4066
+ }
4067
+ },
4068
+ "metadata": {
4069
+ "version": "001",
4070
+ "description": null,
4071
+ "supported_generation_methods": [
4072
+ "generateContent",
4073
+ "countTokens"
4074
+ ]
4075
+ }
4076
+ },
4077
+ {
4078
+ "id": "imagen-3.0-generate-002",
4079
+ "name": "Imagen 3.0 002 model",
4080
+ "provider": "gemini",
4081
+ "family": "imagen3",
4082
+ "created_at": null,
4083
+ "context_window": 480,
4084
+ "max_output_tokens": 8192,
4085
+ "knowledge_cutoff": null,
4086
+ "modalities": {
4087
+ "input": [
4088
+ "text",
4089
+ "image",
4090
+ "pdf"
4091
+ ],
4092
+ "output": [
4093
+ "text"
4094
+ ]
4095
+ },
4096
+ "capabilities": [
4097
+ "streaming"
4098
+ ],
4099
+ "pricing": {},
4100
+ "metadata": {
4101
+ "version": "002",
4102
+ "description": "Vertex served Imagen 3.0 002 model",
4103
+ "supported_generation_methods": [
4104
+ "predict"
4105
+ ]
4106
+ }
4107
+ },
4108
+ {
4109
+ "id": "imagen-4.0-generate-preview-06-06",
4110
+ "name": "Imagen 4 (Preview)",
4111
+ "provider": "gemini",
4112
+ "family": "other",
4113
+ "created_at": null,
4114
+ "context_window": 480,
4115
+ "max_output_tokens": 8192,
4116
+ "knowledge_cutoff": null,
4117
+ "modalities": {
4118
+ "input": [
4119
+ "text",
4120
+ "image",
4121
+ "pdf"
4122
+ ],
4123
+ "output": [
4124
+ "text"
4125
+ ]
4126
+ },
4127
+ "capabilities": [
4128
+ "streaming"
4129
+ ],
4130
+ "pricing": {},
4131
+ "metadata": {
4132
+ "version": "01",
4133
+ "description": "Vertex served Imagen 4.0 model",
4134
+ "supported_generation_methods": [
4135
+ "predict"
4136
+ ]
4137
+ }
4138
+ },
4139
+ {
4140
+ "id": "imagen-4.0-ultra-generate-preview-06-06",
4141
+ "name": "Imagen 4 Ultra (Preview)",
4142
+ "provider": "gemini",
4143
+ "family": "other",
4144
+ "created_at": null,
4145
+ "context_window": 480,
4146
+ "max_output_tokens": 8192,
4147
+ "knowledge_cutoff": null,
4148
+ "modalities": {
4149
+ "input": [
4150
+ "text",
4151
+ "image",
4152
+ "pdf"
4153
+ ],
4154
+ "output": [
4155
+ "text"
4156
+ ]
4157
+ },
4158
+ "capabilities": [
4159
+ "streaming"
4160
+ ],
4161
+ "pricing": {},
4162
+ "metadata": {
4163
+ "version": "01",
4164
+ "description": "Vertex served Imagen 4.0 ultra model",
4165
+ "supported_generation_methods": [
4166
+ "predict"
4167
+ ]
4168
+ }
4169
+ },
4170
+ {
4171
+ "id": "learnlm-2.0-flash-experimental",
4172
+ "name": "LearnLM 2.0 Flash Experimental",
4173
+ "provider": "gemini",
4174
+ "family": "other",
4175
+ "created_at": null,
4176
+ "context_window": 1048576,
4177
+ "max_output_tokens": 32768,
4178
+ "knowledge_cutoff": null,
4179
+ "modalities": {
4180
+ "input": [
4181
+ "text",
4182
+ "image",
4183
+ "pdf"
4184
+ ],
4185
+ "output": [
4186
+ "text"
4187
+ ]
4188
+ },
4189
+ "capabilities": [
4190
+ "streaming",
4191
+ "function_calling",
4192
+ "structured_output",
4193
+ "batch",
4194
+ "caching"
4195
+ ],
4196
+ "pricing": {
4197
+ "text_tokens": {
4198
+ "standard": {
4199
+ "input_per_million": 0.075,
4200
+ "output_per_million": 0.3
4201
+ },
4202
+ "batch": {
4203
+ "input_per_million": 0.0375,
4204
+ "output_per_million": 0.15
4205
+ }
4206
+ }
4207
+ },
4208
+ "metadata": {
4209
+ "version": "2.0",
4210
+ "description": "LearnLM 2.0 Flash Experimental",
4211
+ "supported_generation_methods": [
4212
+ "generateContent",
4213
+ "countTokens"
4214
+ ]
4215
+ }
4216
+ },
4217
+ {
4218
+ "id": "text-embedding-004",
4219
+ "name": "Text Embedding 004",
4220
+ "provider": "gemini",
4221
+ "family": "embedding4",
4222
+ "created_at": null,
4223
+ "context_window": 2048,
4224
+ "max_output_tokens": 1,
4225
+ "knowledge_cutoff": null,
4226
+ "modalities": {
4227
+ "input": [
4228
+ "text"
4229
+ ],
4230
+ "output": [
4231
+ "text",
4232
+ "embeddings"
4233
+ ]
4234
+ },
4235
+ "capabilities": [
4236
+ "streaming",
4237
+ "batch"
4238
+ ],
4239
+ "pricing": {
4240
+ "embeddings": {
4241
+ "standard": {
4242
+ "input_per_million": 0.002
4243
+ }
4244
+ }
4245
+ },
4246
+ "metadata": {
4247
+ "version": "004",
4248
+ "description": "Obtain a distributed representation of a text.",
4249
+ "supported_generation_methods": [
4250
+ "embedContent"
4251
+ ]
4252
+ }
4253
+ },
4254
+ {
4255
+ "id": "veo-2.0-generate-001",
4256
+ "name": "Veo 2",
4257
+ "provider": "gemini",
4258
+ "family": "other",
4259
+ "created_at": null,
4260
+ "context_window": 480,
4261
+ "max_output_tokens": 8192,
4262
+ "knowledge_cutoff": null,
4263
+ "modalities": {
4264
+ "input": [
4265
+ "text"
4266
+ ],
4267
+ "output": [
4268
+ "text"
4269
+ ]
4270
+ },
4271
+ "capabilities": [
4272
+ "streaming"
4273
+ ],
4274
+ "pricing": {
4275
+ "text_tokens": {
4276
+ "standard": {
4277
+ "input_per_million": 0.075,
4278
+ "output_per_million": 0.3
4279
+ },
4280
+ "batch": {
4281
+ "input_per_million": 0.0375,
4282
+ "output_per_million": 0.15
4283
+ }
4284
+ }
4285
+ },
4286
+ "metadata": {
4287
+ "version": "2.0",
4288
+ "description": "Vertex served Veo 2 model. Access to this model requires billing to be enabled on the associated Google Cloud Platform account. Please visit https://console.cloud.google.com/billing to enable it.",
4289
+ "supported_generation_methods": [
4290
+ "predictLongRunning"
4291
+ ]
4292
+ }
4293
+ },
4294
+ {
4295
+ "id": "codestral-2411-rc5",
4296
+ "name": "Codestral",
4297
+ "provider": "mistral",
4298
+ "family": "codestral",
4299
+ "created_at": "2024-11-26 00:00:00 +0100",
4300
+ "context_window": 32768,
4301
+ "max_output_tokens": 8192,
4302
+ "knowledge_cutoff": null,
4303
+ "modalities": {
4304
+ "input": [
4305
+ "text"
4306
+ ],
4307
+ "output": [
4308
+ "text"
4309
+ ]
4310
+ },
4311
+ "capabilities": {
4312
+ "chat": true,
4313
+ "streaming": true,
4314
+ "tools": true,
4315
+ "vision": false,
4316
+ "json_mode": true
4317
+ },
4318
+ "pricing": {},
4319
+ "metadata": {
4320
+ "object": "model",
4321
+ "owned_by": "mistralai"
4322
+ }
4323
+ },
4324
+ {
4325
+ "id": "codestral-2412",
4326
+ "name": "Codestral",
4327
+ "provider": "mistral",
4328
+ "family": "codestral",
4329
+ "created_at": "2024-12-17 00:00:00 +0100",
4330
+ "context_window": 32768,
4331
+ "max_output_tokens": 8192,
4332
+ "knowledge_cutoff": null,
4333
+ "modalities": {
4334
+ "input": [
4335
+ "text"
4336
+ ],
4337
+ "output": [
4338
+ "text"
4339
+ ]
4340
+ },
4341
+ "capabilities": {
4342
+ "chat": true,
4343
+ "streaming": true,
4344
+ "tools": true,
4345
+ "vision": false,
4346
+ "json_mode": true
4347
+ },
4348
+ "pricing": {},
4349
+ "metadata": {
4350
+ "object": "model",
4351
+ "owned_by": "mistralai"
4352
+ }
4353
+ },
4354
+ {
4355
+ "id": "codestral-2501",
4356
+ "name": "Codestral",
4357
+ "provider": "mistral",
4358
+ "family": "codestral",
4359
+ "created_at": "2025-01-14 00:00:00 +0100",
4360
+ "context_window": 32768,
4361
+ "max_output_tokens": 8192,
4362
+ "knowledge_cutoff": null,
4363
+ "modalities": {
4364
+ "input": [
4365
+ "text"
4366
+ ],
4367
+ "output": [
4368
+ "text"
4369
+ ]
4370
+ },
4371
+ "capabilities": {
4372
+ "chat": true,
4373
+ "streaming": true,
4374
+ "tools": true,
4375
+ "vision": false,
4376
+ "json_mode": true
4377
+ },
4378
+ "pricing": {},
4379
+ "metadata": {
4380
+ "object": "model",
4381
+ "owned_by": "mistralai"
4382
+ }
4383
+ },
4384
+ {
4385
+ "id": "codestral-2508",
4386
+ "name": "Codestral",
4387
+ "provider": "mistral",
4388
+ "family": "codestral",
4389
+ "created_at": "2025-08-30 00:00:00 +0200",
4390
+ "context_window": 32768,
4391
+ "max_output_tokens": 8192,
4392
+ "knowledge_cutoff": null,
4393
+ "modalities": {
4394
+ "input": [
4395
+ "text"
4396
+ ],
4397
+ "output": [
4398
+ "text"
4399
+ ]
4400
+ },
4401
+ "capabilities": {
4402
+ "chat": true,
4403
+ "streaming": true,
4404
+ "tools": true,
4405
+ "vision": false,
4406
+ "json_mode": true
4407
+ },
4408
+ "pricing": {},
4409
+ "metadata": {
4410
+ "object": "model",
4411
+ "owned_by": "mistralai"
4412
+ }
4413
+ },
4414
+ {
4415
+ "id": "codestral-embed",
4416
+ "name": "Codestral",
4417
+ "provider": "mistral",
4418
+ "family": "codestral",
4419
+ "created_at": "2025-05-21 00:00:00 +0200",
4420
+ "context_window": 32768,
4421
+ "max_output_tokens": 8192,
4422
+ "knowledge_cutoff": null,
4423
+ "modalities": {
4424
+ "input": [
4425
+ "text"
4426
+ ],
4427
+ "output": [
4428
+ "embedding"
4429
+ ]
4430
+ },
4431
+ "capabilities": {
4432
+ "embeddings": true
4433
+ },
4434
+ "pricing": {},
4435
+ "metadata": {
4436
+ "object": "model",
4437
+ "owned_by": "mistralai"
4438
+ }
4439
+ },
4440
+ {
4441
+ "id": "codestral-embed-2505",
4442
+ "name": "Codestral",
4443
+ "provider": "mistral",
4444
+ "family": "codestral",
4445
+ "created_at": "2025-05-21 00:00:00 +0200",
4446
+ "context_window": 32768,
4447
+ "max_output_tokens": 8192,
4448
+ "knowledge_cutoff": null,
4449
+ "modalities": {
4450
+ "input": [
4451
+ "text"
4452
+ ],
4453
+ "output": [
4454
+ "embedding"
4455
+ ]
4456
+ },
4457
+ "capabilities": {
4458
+ "embeddings": true
4459
+ },
4460
+ "pricing": {},
4461
+ "metadata": {
4462
+ "object": "model",
4463
+ "owned_by": "mistralai"
4464
+ }
4465
+ },
4466
+ {
4467
+ "id": "codestral-latest",
4468
+ "name": "Codestral",
4469
+ "provider": "mistral",
4470
+ "family": "codestral",
4471
+ "created_at": "2025-08-30 00:00:00 +0200",
4472
+ "context_window": 32768,
4473
+ "max_output_tokens": 8192,
4474
+ "knowledge_cutoff": null,
4475
+ "modalities": {
4476
+ "input": [
4477
+ "text"
4478
+ ],
4479
+ "output": [
4480
+ "text"
4481
+ ]
4482
+ },
4483
+ "capabilities": {
4484
+ "chat": true,
4485
+ "streaming": true,
4486
+ "tools": true,
4487
+ "vision": false,
4488
+ "json_mode": true
4489
+ },
4490
+ "pricing": {},
4491
+ "metadata": {
4492
+ "object": "model",
4493
+ "owned_by": "mistralai"
4494
+ }
4495
+ },
4496
+ {
4497
+ "id": "devstral-medium-2507",
4498
+ "name": "Devstral Medium 2507",
4499
+ "provider": "mistral",
4500
+ "family": "mistral",
4501
+ "created_at": "2025-07-09 00:00:00 +0200",
4502
+ "context_window": 32768,
4503
+ "max_output_tokens": 8192,
4504
+ "knowledge_cutoff": null,
4505
+ "modalities": {
4506
+ "input": [
4507
+ "text"
4508
+ ],
4509
+ "output": [
4510
+ "text"
4511
+ ]
4512
+ },
4513
+ "capabilities": {
4514
+ "chat": true,
4515
+ "streaming": true,
4516
+ "tools": true,
4517
+ "vision": false,
4518
+ "json_mode": true
4519
+ },
4520
+ "pricing": {},
4521
+ "metadata": {
4522
+ "object": "model",
4523
+ "owned_by": "mistralai"
4524
+ }
4525
+ },
4526
+ {
4527
+ "id": "devstral-medium-latest",
4528
+ "name": "Devstral Medium Latest",
4529
+ "provider": "mistral",
4530
+ "family": "mistral",
4531
+ "created_at": "2025-07-09 00:00:00 +0200",
4532
+ "context_window": 32768,
4533
+ "max_output_tokens": 8192,
4534
+ "knowledge_cutoff": null,
4535
+ "modalities": {
4536
+ "input": [
4537
+ "text"
4538
+ ],
4539
+ "output": [
4540
+ "text"
4541
+ ]
4542
+ },
4543
+ "capabilities": {
4544
+ "chat": true,
4545
+ "streaming": true,
4546
+ "tools": true,
4547
+ "vision": false,
4548
+ "json_mode": true
4549
+ },
4550
+ "pricing": {},
4551
+ "metadata": {
4552
+ "object": "model",
4553
+ "owned_by": "mistralai"
4554
+ }
4555
+ },
4556
+ {
4557
+ "id": "devstral-small-2505",
4558
+ "name": "Devstral Small 2505",
4559
+ "provider": "mistral",
4560
+ "family": "mistral",
4561
+ "created_at": "2025-05-28 00:00:00 +0200",
4562
+ "context_window": 32768,
4563
+ "max_output_tokens": 8192,
4564
+ "knowledge_cutoff": null,
4565
+ "modalities": {
4566
+ "input": [
4567
+ "text"
4568
+ ],
4569
+ "output": [
4570
+ "text"
4571
+ ]
4572
+ },
4573
+ "capabilities": {
4574
+ "chat": true,
4575
+ "streaming": true,
4576
+ "tools": true,
4577
+ "vision": false,
4578
+ "json_mode": true
4579
+ },
4580
+ "pricing": {},
4581
+ "metadata": {
4582
+ "object": "model",
4583
+ "owned_by": "mistralai"
4584
+ }
4585
+ },
4586
+ {
4587
+ "id": "devstral-small-2507",
4588
+ "name": "Devstral Small 2507",
4589
+ "provider": "mistral",
4590
+ "family": "mistral",
4591
+ "created_at": "2025-07-09 00:00:00 +0200",
4592
+ "context_window": 32768,
4593
+ "max_output_tokens": 8192,
4594
+ "knowledge_cutoff": null,
4595
+ "modalities": {
4596
+ "input": [
4597
+ "text"
4598
+ ],
4599
+ "output": [
4600
+ "text"
4601
+ ]
4602
+ },
4603
+ "capabilities": {
4604
+ "chat": true,
4605
+ "streaming": true,
4606
+ "tools": true,
4607
+ "vision": false,
4608
+ "json_mode": true
4609
+ },
4610
+ "pricing": {},
4611
+ "metadata": {
4612
+ "object": "model",
4613
+ "owned_by": "mistralai"
4614
+ }
4615
+ },
4616
+ {
4617
+ "id": "devstral-small-latest",
4618
+ "name": "Devstral Small Latest",
4619
+ "provider": "mistral",
4620
+ "family": "mistral",
4621
+ "created_at": "2025-07-09 00:00:00 +0200",
4622
+ "context_window": 32768,
4623
+ "max_output_tokens": 8192,
4624
+ "knowledge_cutoff": null,
4625
+ "modalities": {
4626
+ "input": [
4627
+ "text"
4628
+ ],
4629
+ "output": [
4630
+ "text"
4631
+ ]
4632
+ },
4633
+ "capabilities": {
4634
+ "chat": true,
4635
+ "streaming": true,
4636
+ "tools": true,
4637
+ "vision": false,
4638
+ "json_mode": true
4639
+ },
4640
+ "pricing": {},
4641
+ "metadata": {
4642
+ "object": "model",
4643
+ "owned_by": "mistralai"
4644
+ }
4645
+ },
4646
+ {
4647
+ "id": "magistral-medium-2506",
4648
+ "name": "Magistral Medium 2506",
4649
+ "provider": "mistral",
4650
+ "family": "mistral",
4651
+ "created_at": "2025-06-10 00:00:00 +0200",
4652
+ "context_window": 32768,
4653
+ "max_output_tokens": 8192,
4654
+ "knowledge_cutoff": null,
4655
+ "modalities": {
4656
+ "input": [
4657
+ "text"
4658
+ ],
4659
+ "output": [
4660
+ "text"
4661
+ ]
4662
+ },
4663
+ "capabilities": {
4664
+ "chat": true,
4665
+ "streaming": true,
4666
+ "tools": true,
4667
+ "vision": false,
4668
+ "json_mode": true
4669
+ },
4670
+ "pricing": {},
4671
+ "metadata": {
4672
+ "object": "model",
4673
+ "owned_by": "mistralai"
4674
+ }
4675
+ },
4676
+ {
4677
+ "id": "magistral-medium-2507",
4678
+ "name": "Magistral Medium 2507",
4679
+ "provider": "mistral",
4680
+ "family": "mistral",
4681
+ "created_at": null,
4682
+ "context_window": 32768,
4683
+ "max_output_tokens": 8192,
4684
+ "knowledge_cutoff": null,
4685
+ "modalities": {
4686
+ "input": [
4687
+ "text"
4688
+ ],
4689
+ "output": [
4690
+ "text"
4691
+ ]
4692
+ },
4693
+ "capabilities": {
4694
+ "chat": true,
4695
+ "streaming": true,
4696
+ "tools": true,
4697
+ "vision": false,
4698
+ "json_mode": true
4699
+ },
4700
+ "pricing": {},
4701
+ "metadata": {
4702
+ "object": "model",
4703
+ "owned_by": "mistralai"
4704
+ }
4705
+ },
4706
+ {
4707
+ "id": "magistral-medium-latest",
4708
+ "name": "Magistral Medium Latest",
4709
+ "provider": "mistral",
4710
+ "family": "mistral",
4711
+ "created_at": "2025-06-10 00:00:00 +0200",
4712
+ "context_window": 32768,
4713
+ "max_output_tokens": 8192,
4714
+ "knowledge_cutoff": null,
4715
+ "modalities": {
4716
+ "input": [
4717
+ "text"
4718
+ ],
4719
+ "output": [
4720
+ "text"
4721
+ ]
4722
+ },
4723
+ "capabilities": {
4724
+ "chat": true,
4725
+ "streaming": true,
4726
+ "tools": true,
4727
+ "vision": false,
4728
+ "json_mode": true
4729
+ },
4730
+ "pricing": {},
4731
+ "metadata": {
4732
+ "object": "model",
4733
+ "owned_by": "mistralai"
4734
+ }
4735
+ },
4736
+ {
4737
+ "id": "magistral-small-2506",
4738
+ "name": "Magistral Small 2506",
4739
+ "provider": "mistral",
4740
+ "family": "mistral",
4741
+ "created_at": null,
4742
+ "context_window": 32768,
4743
+ "max_output_tokens": 8192,
4744
+ "knowledge_cutoff": null,
4745
+ "modalities": {
4746
+ "input": [
4747
+ "text"
4748
+ ],
4749
+ "output": [
4750
+ "text"
4751
+ ]
4752
+ },
4753
+ "capabilities": {
4754
+ "chat": true,
4755
+ "streaming": true,
4756
+ "tools": true,
4757
+ "vision": false,
4758
+ "json_mode": true
4759
+ },
4760
+ "pricing": {},
4761
+ "metadata": {
4762
+ "object": "model",
4763
+ "owned_by": "mistralai"
4764
+ }
4765
+ },
4766
+ {
4767
+ "id": "magistral-small-2507",
4768
+ "name": "Magistral Small 2507",
4769
+ "provider": "mistral",
4770
+ "family": "mistral",
4771
+ "created_at": null,
4772
+ "context_window": 32768,
4773
+ "max_output_tokens": 8192,
4774
+ "knowledge_cutoff": null,
4775
+ "modalities": {
4776
+ "input": [
4777
+ "text"
4778
+ ],
4779
+ "output": [
4780
+ "text"
4781
+ ]
4782
+ },
4783
+ "capabilities": {
4784
+ "chat": true,
4785
+ "streaming": true,
4786
+ "tools": true,
4787
+ "vision": false,
4788
+ "json_mode": true
4789
+ },
4790
+ "pricing": {},
4791
+ "metadata": {
4792
+ "object": "model",
4793
+ "owned_by": "mistralai"
4794
+ }
4795
+ },
4796
+ {
4797
+ "id": "magistral-small-latest",
4798
+ "name": "Magistral Small Latest",
4799
+ "provider": "mistral",
4800
+ "family": "mistral",
4801
+ "created_at": null,
4802
+ "context_window": 32768,
4803
+ "max_output_tokens": 8192,
4804
+ "knowledge_cutoff": null,
4805
+ "modalities": {
4806
+ "input": [
4807
+ "text"
4808
+ ],
4809
+ "output": [
4810
+ "text"
4811
+ ]
4812
+ },
4813
+ "capabilities": {
4814
+ "chat": true,
4815
+ "streaming": true,
4816
+ "tools": true,
4817
+ "vision": false,
4818
+ "json_mode": true
4819
+ },
4820
+ "pricing": {},
4821
+ "metadata": {
4822
+ "object": "model",
4823
+ "owned_by": "mistralai"
4824
+ }
4825
+ },
4826
+ {
4827
+ "id": "ministral-3b-2410",
4828
+ "name": "Ministral 3B",
4829
+ "provider": "mistral",
4830
+ "family": "ministral",
4831
+ "created_at": "2024-10-16 00:00:00 +0200",
4832
+ "context_window": 32768,
4833
+ "max_output_tokens": 8192,
4834
+ "knowledge_cutoff": null,
4835
+ "modalities": {
4836
+ "input": [
4837
+ "text"
4838
+ ],
4839
+ "output": [
4840
+ "text"
4841
+ ]
4842
+ },
4843
+ "capabilities": {
4844
+ "chat": true,
4845
+ "streaming": true,
4846
+ "tools": true,
4847
+ "vision": false,
4848
+ "json_mode": true
4849
+ },
4850
+ "pricing": {},
4851
+ "metadata": {
4852
+ "object": "model",
4853
+ "owned_by": "mistralai"
4854
+ }
4855
+ },
4856
+ {
4857
+ "id": "ministral-3b-latest",
4858
+ "name": "Ministral 3B",
4859
+ "provider": "mistral",
4860
+ "family": "ministral",
4861
+ "created_at": "2024-10-16 00:00:00 +0200",
4862
+ "context_window": 32768,
4863
+ "max_output_tokens": 8192,
4864
+ "knowledge_cutoff": null,
4865
+ "modalities": {
4866
+ "input": [
4867
+ "text"
4868
+ ],
4869
+ "output": [
4870
+ "text"
4871
+ ]
4872
+ },
4873
+ "capabilities": {
4874
+ "chat": true,
4875
+ "streaming": true,
4876
+ "tools": true,
4877
+ "vision": false,
4878
+ "json_mode": true
4879
+ },
4880
+ "pricing": {},
4881
+ "metadata": {
4882
+ "object": "model",
4883
+ "owned_by": "mistralai"
4884
+ }
4885
+ },
4886
+ {
4887
+ "id": "ministral-8b-2410",
4888
+ "name": "Ministral 8B",
4889
+ "provider": "mistral",
4890
+ "family": "ministral",
4891
+ "created_at": "2024-10-16 00:00:00 +0200",
4892
+ "context_window": 32768,
4893
+ "max_output_tokens": 8192,
4894
+ "knowledge_cutoff": null,
4895
+ "modalities": {
4896
+ "input": [
4897
+ "text"
4898
+ ],
4899
+ "output": [
4900
+ "text"
4901
+ ]
4902
+ },
4903
+ "capabilities": {
4904
+ "chat": true,
4905
+ "streaming": true,
4906
+ "tools": true,
4907
+ "vision": false,
4908
+ "json_mode": true
4909
+ },
4910
+ "pricing": {},
4911
+ "metadata": {
4912
+ "object": "model",
4913
+ "owned_by": "mistralai"
4914
+ }
4915
+ },
4916
+ {
4917
+ "id": "ministral-8b-latest",
4918
+ "name": "Ministral 8B",
4919
+ "provider": "mistral",
4920
+ "family": "ministral",
4921
+ "created_at": "2024-10-16 00:00:00 +0200",
4922
+ "context_window": 32768,
4923
+ "max_output_tokens": 8192,
4924
+ "knowledge_cutoff": null,
4925
+ "modalities": {
4926
+ "input": [
4927
+ "text"
4928
+ ],
4929
+ "output": [
4930
+ "text"
4931
+ ]
4932
+ },
4933
+ "capabilities": {
4934
+ "chat": true,
4935
+ "streaming": true,
4936
+ "tools": true,
4937
+ "vision": false,
4938
+ "json_mode": true
4939
+ },
4940
+ "pricing": {},
4941
+ "metadata": {
4942
+ "object": "model",
4943
+ "owned_by": "mistralai"
4944
+ }
4945
+ },
4946
+ {
4947
+ "id": "mistral-embed",
4948
+ "name": "Mistral Embed",
4949
+ "provider": "mistral",
4950
+ "family": "mistral-embed",
4951
+ "created_at": "2024-01-11 00:00:00 +0100",
4952
+ "context_window": 32768,
4953
+ "max_output_tokens": 8192,
4954
+ "knowledge_cutoff": null,
4955
+ "modalities": {
4956
+ "input": [
4957
+ "text"
4958
+ ],
4959
+ "output": [
4960
+ "embedding"
4961
+ ]
4962
+ },
4963
+ "capabilities": {
4964
+ "embeddings": true
4965
+ },
4966
+ "pricing": {},
4967
+ "metadata": {
4968
+ "object": "model",
4969
+ "owned_by": "mistralai"
4970
+ }
4971
+ },
4972
+ {
4973
+ "id": "mistral-large-2407",
4974
+ "name": "Mistral Large",
4975
+ "provider": "mistral",
4976
+ "family": "mistral-large",
4977
+ "created_at": "2024-07-24 00:00:00 +0200",
4978
+ "context_window": 32768,
4979
+ "max_output_tokens": 8192,
4980
+ "knowledge_cutoff": null,
4981
+ "modalities": {
4982
+ "input": [
4983
+ "text"
4984
+ ],
4985
+ "output": [
4986
+ "text"
4987
+ ]
4988
+ },
4989
+ "capabilities": {
4990
+ "chat": true,
4991
+ "streaming": true,
4992
+ "tools": true,
4993
+ "vision": false,
4994
+ "json_mode": true
4995
+ },
4996
+ "pricing": {},
4997
+ "metadata": {
4998
+ "object": "model",
4999
+ "owned_by": "mistralai"
5000
+ }
5001
+ },
5002
+ {
5003
+ "id": "mistral-large-2411",
5004
+ "name": "Mistral Large",
5005
+ "provider": "mistral",
5006
+ "family": "mistral-large",
5007
+ "created_at": "2024-11-20 00:00:00 +0100",
5008
+ "context_window": 32768,
5009
+ "max_output_tokens": 8192,
5010
+ "knowledge_cutoff": null,
5011
+ "modalities": {
5012
+ "input": [
5013
+ "text"
5014
+ ],
5015
+ "output": [
5016
+ "text"
5017
+ ]
5018
+ },
5019
+ "capabilities": {
5020
+ "chat": true,
5021
+ "streaming": true,
5022
+ "tools": true,
5023
+ "vision": false,
5024
+ "json_mode": true
5025
+ },
5026
+ "pricing": {},
5027
+ "metadata": {
5028
+ "object": "model",
5029
+ "owned_by": "mistralai"
5030
+ }
5031
+ },
5032
+ {
5033
+ "id": "mistral-large-latest",
5034
+ "name": "Mistral Large",
5035
+ "provider": "mistral",
5036
+ "family": "mistral-large",
5037
+ "created_at": "2024-11-20 00:00:00 +0100",
5038
+ "context_window": 32768,
5039
+ "max_output_tokens": 8192,
5040
+ "knowledge_cutoff": null,
5041
+ "modalities": {
5042
+ "input": [
5043
+ "text"
5044
+ ],
5045
+ "output": [
5046
+ "text"
5047
+ ]
5048
+ },
5049
+ "capabilities": {
5050
+ "chat": true,
5051
+ "streaming": true,
5052
+ "tools": true,
5053
+ "vision": false,
5054
+ "json_mode": true
5055
+ },
5056
+ "pricing": {},
5057
+ "metadata": {
5058
+ "object": "model",
5059
+ "owned_by": "mistralai"
5060
+ }
5061
+ },
5062
+ {
5063
+ "id": "mistral-large-pixtral-2411",
5064
+ "name": "Mistral Large",
5065
+ "provider": "mistral",
5066
+ "family": "mistral-large",
5067
+ "created_at": "2024-11-12 00:00:00 +0100",
5068
+ "context_window": 32768,
5069
+ "max_output_tokens": 8192,
5070
+ "knowledge_cutoff": null,
5071
+ "modalities": {
5072
+ "input": [
5073
+ "text",
5074
+ "image"
5075
+ ],
5076
+ "output": [
5077
+ "text"
5078
+ ]
5079
+ },
5080
+ "capabilities": {
5081
+ "chat": true,
5082
+ "streaming": true,
5083
+ "tools": true,
5084
+ "vision": true,
5085
+ "json_mode": true
5086
+ },
5087
+ "pricing": {},
5088
+ "metadata": {
5089
+ "object": "model",
5090
+ "owned_by": "mistralai"
5091
+ }
5092
+ },
5093
+ {
5094
+ "id": "mistral-medium",
5095
+ "name": "Mistral Medium",
5096
+ "provider": "mistral",
5097
+ "family": "mistral-medium",
5098
+ "created_at": "2025-05-06 00:00:00 +0200",
5099
+ "context_window": 32768,
5100
+ "max_output_tokens": 8192,
5101
+ "knowledge_cutoff": null,
5102
+ "modalities": {
5103
+ "input": [
5104
+ "text"
5105
+ ],
5106
+ "output": [
5107
+ "text"
5108
+ ]
5109
+ },
5110
+ "capabilities": {
5111
+ "chat": true,
5112
+ "streaming": true,
5113
+ "tools": true,
5114
+ "vision": false,
5115
+ "json_mode": true
5116
+ },
5117
+ "pricing": {},
5118
+ "metadata": {
5119
+ "object": "model",
5120
+ "owned_by": "mistralai"
5121
+ }
5122
+ },
5123
+ {
5124
+ "id": "mistral-medium-2505",
5125
+ "name": "Mistral Medium",
5126
+ "provider": "mistral",
5127
+ "family": "mistral-medium",
5128
+ "created_at": "2025-05-06 00:00:00 +0200",
5129
+ "context_window": 32768,
5130
+ "max_output_tokens": 8192,
5131
+ "knowledge_cutoff": null,
5132
+ "modalities": {
5133
+ "input": [
5134
+ "text"
5135
+ ],
5136
+ "output": [
5137
+ "text"
5138
+ ]
5139
+ },
5140
+ "capabilities": {
5141
+ "chat": true,
5142
+ "streaming": true,
5143
+ "tools": true,
5144
+ "vision": false,
5145
+ "json_mode": true
5146
+ },
5147
+ "pricing": {},
5148
+ "metadata": {
5149
+ "object": "model",
5150
+ "owned_by": "mistralai"
5151
+ }
5152
+ },
5153
+ {
5154
+ "id": "mistral-medium-latest",
5155
+ "name": "Mistral Medium",
5156
+ "provider": "mistral",
5157
+ "family": "mistral-medium",
5158
+ "created_at": "2025-05-06 00:00:00 +0200",
5159
+ "context_window": 32768,
5160
+ "max_output_tokens": 8192,
5161
+ "knowledge_cutoff": null,
5162
+ "modalities": {
5163
+ "input": [
5164
+ "text"
5165
+ ],
5166
+ "output": [
5167
+ "text"
5168
+ ]
5169
+ },
5170
+ "capabilities": {
5171
+ "chat": true,
5172
+ "streaming": true,
5173
+ "tools": true,
5174
+ "vision": false,
5175
+ "json_mode": true
5176
+ },
5177
+ "pricing": {},
5178
+ "metadata": {
5179
+ "object": "model",
5180
+ "owned_by": "mistralai"
5181
+ }
5182
+ },
5183
+ {
5184
+ "id": "mistral-moderation-2411",
5185
+ "name": "Mistral Moderation",
5186
+ "provider": "mistral",
5187
+ "family": "mistral-moderation",
5188
+ "created_at": "2024-11-26 00:00:00 +0100",
5189
+ "context_window": 32768,
5190
+ "max_output_tokens": 8192,
5191
+ "knowledge_cutoff": null,
5192
+ "modalities": {
5193
+ "input": [
5194
+ "text"
5195
+ ],
5196
+ "output": [
5197
+ "text"
5198
+ ]
5199
+ },
5200
+ "capabilities": {
5201
+ "moderation": true
5202
+ },
5203
+ "pricing": {},
5204
+ "metadata": {
5205
+ "object": "model",
5206
+ "owned_by": "mistralai"
5207
+ }
5208
+ },
5209
+ {
5210
+ "id": "mistral-moderation-latest",
5211
+ "name": "Mistral Moderation",
5212
+ "provider": "mistral",
5213
+ "family": "mistral-moderation",
5214
+ "created_at": "2024-11-26 00:00:00 +0100",
5215
+ "context_window": 32768,
5216
+ "max_output_tokens": 8192,
5217
+ "knowledge_cutoff": null,
5218
+ "modalities": {
5219
+ "input": [
5220
+ "text"
5221
+ ],
5222
+ "output": [
5223
+ "text"
5224
+ ]
5225
+ },
5226
+ "capabilities": {
5227
+ "moderation": true
5228
+ },
5229
+ "pricing": {},
5230
+ "metadata": {
5231
+ "object": "model",
5232
+ "owned_by": "mistralai"
5233
+ }
5234
+ },
5235
+ {
5236
+ "id": "mistral-ocr-2503",
5237
+ "name": "Mistral Ocr 2503",
5238
+ "provider": "mistral",
5239
+ "family": "mistral",
5240
+ "created_at": "2025-03-21 00:00:00 +0100",
5241
+ "context_window": 32768,
5242
+ "max_output_tokens": 8192,
5243
+ "knowledge_cutoff": null,
5244
+ "modalities": {
5245
+ "input": [
5246
+ "text"
5247
+ ],
5248
+ "output": [
5249
+ "text"
5250
+ ]
5251
+ },
5252
+ "capabilities": {
5253
+ "chat": true,
5254
+ "streaming": true,
5255
+ "tools": true,
5256
+ "vision": false,
5257
+ "json_mode": true
5258
+ },
5259
+ "pricing": {},
5260
+ "metadata": {
5261
+ "object": "model",
5262
+ "owned_by": "mistralai"
5263
+ }
5264
+ },
5265
+ {
5266
+ "id": "mistral-ocr-2505",
5267
+ "name": "Mistral Ocr 2505",
5268
+ "provider": "mistral",
5269
+ "family": "mistral",
5270
+ "created_at": "2025-05-23 00:00:00 +0200",
5271
+ "context_window": 32768,
5272
+ "max_output_tokens": 8192,
5273
+ "knowledge_cutoff": null,
5274
+ "modalities": {
5275
+ "input": [
5276
+ "text"
5277
+ ],
5278
+ "output": [
5279
+ "text"
5280
+ ]
5281
+ },
5282
+ "capabilities": {
5283
+ "chat": true,
5284
+ "streaming": true,
5285
+ "tools": true,
5286
+ "vision": false,
5287
+ "json_mode": true
5288
+ },
5289
+ "pricing": {},
5290
+ "metadata": {
5291
+ "object": "model",
5292
+ "owned_by": "mistralai"
5293
+ }
5294
+ },
5295
+ {
5296
+ "id": "mistral-ocr-latest",
5297
+ "name": "Mistral Ocr Latest",
5298
+ "provider": "mistral",
5299
+ "family": "mistral",
5300
+ "created_at": "2025-05-23 00:00:00 +0200",
5301
+ "context_window": 32768,
5302
+ "max_output_tokens": 8192,
5303
+ "knowledge_cutoff": null,
5304
+ "modalities": {
5305
+ "input": [
5306
+ "text"
5307
+ ],
5308
+ "output": [
5309
+ "text"
5310
+ ]
5311
+ },
5312
+ "capabilities": {
5313
+ "chat": true,
5314
+ "streaming": true,
5315
+ "tools": true,
5316
+ "vision": false,
5317
+ "json_mode": true
5318
+ },
5319
+ "pricing": {},
5320
+ "metadata": {
5321
+ "object": "model",
5322
+ "owned_by": "mistralai"
5323
+ }
5324
+ },
5325
+ {
5326
+ "id": "mistral-saba-2502",
5327
+ "name": "Mistral Saba 2502",
5328
+ "provider": "mistral",
5329
+ "family": "mistral",
5330
+ "created_at": "2025-02-18 00:00:00 +0100",
5331
+ "context_window": 32768,
5332
+ "max_output_tokens": 8192,
5333
+ "knowledge_cutoff": null,
5334
+ "modalities": {
5335
+ "input": [
5336
+ "text"
5337
+ ],
5338
+ "output": [
5339
+ "text"
5340
+ ]
5341
+ },
5342
+ "capabilities": {
5343
+ "chat": true,
5344
+ "streaming": true,
5345
+ "tools": true,
5346
+ "vision": false,
5347
+ "json_mode": true
5348
+ },
5349
+ "pricing": {},
5350
+ "metadata": {
5351
+ "object": "model",
5352
+ "owned_by": "mistralai"
5353
+ }
5354
+ },
5355
+ {
5356
+ "id": "mistral-saba-latest",
5357
+ "name": "Mistral Saba Latest",
5358
+ "provider": "mistral",
5359
+ "family": "mistral",
5360
+ "created_at": "2025-02-18 00:00:00 +0100",
5361
+ "context_window": 32768,
5362
+ "max_output_tokens": 8192,
5363
+ "knowledge_cutoff": null,
5364
+ "modalities": {
5365
+ "input": [
5366
+ "text"
5367
+ ],
5368
+ "output": [
5369
+ "text"
5370
+ ]
5371
+ },
5372
+ "capabilities": {
5373
+ "chat": true,
5374
+ "streaming": true,
5375
+ "tools": true,
5376
+ "vision": false,
5377
+ "json_mode": true
5378
+ },
5379
+ "pricing": {},
5380
+ "metadata": {
5381
+ "object": "model",
5382
+ "owned_by": "mistralai"
5383
+ }
5384
+ },
5385
+ {
5386
+ "id": "mistral-small",
5387
+ "name": "Mistral Small",
5388
+ "provider": "mistral",
5389
+ "family": "mistral-small",
5390
+ "created_at": "2023-12-11 00:00:00 +0100",
5391
+ "context_window": 32768,
5392
+ "max_output_tokens": 8192,
5393
+ "knowledge_cutoff": null,
5394
+ "modalities": {
5395
+ "input": [
5396
+ "text"
5397
+ ],
5398
+ "output": [
5399
+ "text"
5400
+ ]
5401
+ },
5402
+ "capabilities": {
5403
+ "chat": true,
5404
+ "streaming": true,
5405
+ "tools": true,
5406
+ "vision": false,
5407
+ "json_mode": true
5408
+ },
5409
+ "pricing": {},
5410
+ "metadata": {
5411
+ "object": "model",
5412
+ "owned_by": "mistralai"
5413
+ }
5414
+ },
5415
+ {
5416
+ "id": "mistral-small-2312",
5417
+ "name": "Mistral Small",
5418
+ "provider": "mistral",
5419
+ "family": "mistral-small",
5420
+ "created_at": "2023-12-11 00:00:00 +0100",
5421
+ "context_window": 32768,
5422
+ "max_output_tokens": 8192,
5423
+ "knowledge_cutoff": null,
5424
+ "modalities": {
5425
+ "input": [
5426
+ "text"
5427
+ ],
5428
+ "output": [
5429
+ "text"
5430
+ ]
5431
+ },
5432
+ "capabilities": {
5433
+ "chat": true,
5434
+ "streaming": true,
5435
+ "tools": true,
5436
+ "vision": false,
5437
+ "json_mode": true
5438
+ },
5439
+ "pricing": {},
5440
+ "metadata": {
5441
+ "object": "model",
5442
+ "owned_by": "mistralai"
5443
+ }
5444
+ },
5445
+ {
5446
+ "id": "mistral-small-2409",
5447
+ "name": "Mistral Small",
5448
+ "provider": "mistral",
5449
+ "family": "mistral-small",
5450
+ "created_at": "2024-09-18 00:00:00 +0200",
5451
+ "context_window": 32768,
5452
+ "max_output_tokens": 8192,
5453
+ "knowledge_cutoff": null,
5454
+ "modalities": {
5455
+ "input": [
5456
+ "text"
5457
+ ],
5458
+ "output": [
5459
+ "text"
5460
+ ]
5461
+ },
5462
+ "capabilities": {
5463
+ "chat": true,
5464
+ "streaming": true,
5465
+ "tools": true,
5466
+ "vision": false,
5467
+ "json_mode": true
5468
+ },
5469
+ "pricing": {},
5470
+ "metadata": {
5471
+ "object": "model",
5472
+ "owned_by": "mistralai"
5473
+ }
5474
+ },
5475
+ {
5476
+ "id": "mistral-small-2501",
5477
+ "name": "Mistral Small",
5478
+ "provider": "mistral",
5479
+ "family": "mistral-small",
5480
+ "created_at": "2025-01-13 00:00:00 +0100",
5481
+ "context_window": 32768,
5482
+ "max_output_tokens": 8192,
5483
+ "knowledge_cutoff": null,
5484
+ "modalities": {
5485
+ "input": [
5486
+ "text"
5487
+ ],
5488
+ "output": [
5489
+ "text"
5490
+ ]
5491
+ },
5492
+ "capabilities": {
5493
+ "chat": true,
5494
+ "streaming": true,
5495
+ "tools": true,
5496
+ "vision": false,
5497
+ "json_mode": true
5498
+ },
5499
+ "pricing": {},
5500
+ "metadata": {
5501
+ "object": "model",
5502
+ "owned_by": "mistralai"
5503
+ }
5504
+ },
5505
+ {
5506
+ "id": "mistral-small-2503",
5507
+ "name": "Mistral Small",
5508
+ "provider": "mistral",
5509
+ "family": "mistral-small",
5510
+ "created_at": "2025-03-03 00:00:00 +0100",
5511
+ "context_window": 32768,
5512
+ "max_output_tokens": 8192,
5513
+ "knowledge_cutoff": null,
5514
+ "modalities": {
5515
+ "input": [
5516
+ "text"
5517
+ ],
5518
+ "output": [
5519
+ "text"
5520
+ ]
5521
+ },
5522
+ "capabilities": {
5523
+ "chat": true,
5524
+ "streaming": true,
5525
+ "tools": true,
5526
+ "vision": false,
5527
+ "json_mode": true
5528
+ },
5529
+ "pricing": {},
5530
+ "metadata": {
5531
+ "object": "model",
5532
+ "owned_by": "mistralai"
5533
+ }
5534
+ },
5535
+ {
5536
+ "id": "mistral-small-2506",
5537
+ "name": "Mistral Small",
5538
+ "provider": "mistral",
5539
+ "family": "mistral-small",
5540
+ "created_at": "2025-06-10 00:00:00 +0200",
5541
+ "context_window": 32768,
5542
+ "max_output_tokens": 8192,
5543
+ "knowledge_cutoff": null,
5544
+ "modalities": {
5545
+ "input": [
5546
+ "text"
5547
+ ],
5548
+ "output": [
5549
+ "text"
5550
+ ]
5551
+ },
5552
+ "capabilities": {
5553
+ "chat": true,
5554
+ "streaming": true,
5555
+ "tools": true,
5556
+ "vision": false,
5557
+ "json_mode": true
5558
+ },
5559
+ "pricing": {},
5560
+ "metadata": {
5561
+ "object": "model",
5562
+ "owned_by": "mistralai"
5563
+ }
5564
+ },
5565
+ {
5566
+ "id": "mistral-small-latest",
5567
+ "name": "Mistral Small",
5568
+ "provider": "mistral",
5569
+ "family": "mistral-small",
5570
+ "created_at": "2025-06-10 00:00:00 +0200",
5571
+ "context_window": 32768,
5572
+ "max_output_tokens": 8192,
5573
+ "knowledge_cutoff": null,
5574
+ "modalities": {
5575
+ "input": [
5576
+ "text"
5577
+ ],
5578
+ "output": [
5579
+ "text"
5580
+ ]
5581
+ },
5582
+ "capabilities": {
5583
+ "chat": true,
5584
+ "streaming": true,
5585
+ "tools": true,
5586
+ "vision": false,
5587
+ "json_mode": true
5588
+ },
5589
+ "pricing": {},
5590
+ "metadata": {
5591
+ "object": "model",
5592
+ "owned_by": "mistralai"
5593
+ }
5594
+ },
5595
+ {
5596
+ "id": "mistral-tiny",
5597
+ "name": "Mistral Tiny",
5598
+ "provider": "mistral",
5599
+ "family": "mistral",
5600
+ "created_at": "2023-09-27 00:00:00 +0200",
5601
+ "context_window": 32768,
5602
+ "max_output_tokens": 8192,
5603
+ "knowledge_cutoff": null,
5604
+ "modalities": {
5605
+ "input": [
5606
+ "text"
5607
+ ],
5608
+ "output": [
5609
+ "text"
5610
+ ]
5611
+ },
5612
+ "capabilities": {
5613
+ "chat": true,
5614
+ "streaming": true,
5615
+ "tools": true,
5616
+ "vision": false,
5617
+ "json_mode": true
5618
+ },
5619
+ "pricing": {},
5620
+ "metadata": {
5621
+ "object": "model",
5622
+ "owned_by": "mistralai"
5623
+ }
5624
+ },
5625
+ {
5626
+ "id": "mistral-tiny-2312",
5627
+ "name": "Mistral Tiny 2312",
5628
+ "provider": "mistral",
5629
+ "family": "mistral",
5630
+ "created_at": "2023-12-11 00:00:00 +0100",
5631
+ "context_window": 32768,
5632
+ "max_output_tokens": 8192,
5633
+ "knowledge_cutoff": null,
5634
+ "modalities": {
5635
+ "input": [
5636
+ "text"
5637
+ ],
5638
+ "output": [
5639
+ "text"
5640
+ ]
5641
+ },
5642
+ "capabilities": {
5643
+ "chat": true,
5644
+ "streaming": true,
5645
+ "tools": true,
5646
+ "vision": false,
5647
+ "json_mode": true
5648
+ },
5649
+ "pricing": {},
5650
+ "metadata": {
5651
+ "object": "model",
5652
+ "owned_by": "mistralai"
5653
+ }
5654
+ },
5655
+ {
5656
+ "id": "mistral-tiny-2407",
5657
+ "name": "Mistral Tiny 2407",
5658
+ "provider": "mistral",
5659
+ "family": "mistral",
5660
+ "created_at": "2024-07-18 00:00:00 +0200",
5661
+ "context_window": 32768,
5662
+ "max_output_tokens": 8192,
5663
+ "knowledge_cutoff": null,
5664
+ "modalities": {
5665
+ "input": [
5666
+ "text"
5667
+ ],
5668
+ "output": [
5669
+ "text"
5670
+ ]
5671
+ },
5672
+ "capabilities": {
5673
+ "chat": true,
5674
+ "streaming": true,
5675
+ "tools": true,
5676
+ "vision": false,
5677
+ "json_mode": true
5678
+ },
5679
+ "pricing": {},
5680
+ "metadata": {
5681
+ "object": "model",
5682
+ "owned_by": "mistralai"
5683
+ }
5684
+ },
5685
+ {
5686
+ "id": "mistral-tiny-latest",
5687
+ "name": "Mistral Tiny Latest",
5688
+ "provider": "mistral",
5689
+ "family": "mistral",
5690
+ "created_at": "2024-07-18 00:00:00 +0200",
5691
+ "context_window": 32768,
5692
+ "max_output_tokens": 8192,
5693
+ "knowledge_cutoff": null,
5694
+ "modalities": {
5695
+ "input": [
5696
+ "text"
5697
+ ],
5698
+ "output": [
5699
+ "text"
5700
+ ]
5701
+ },
5702
+ "capabilities": {
5703
+ "chat": true,
5704
+ "streaming": true,
5705
+ "tools": true,
5706
+ "vision": false,
5707
+ "json_mode": true
5708
+ },
5709
+ "pricing": {},
5710
+ "metadata": {
5711
+ "object": "model",
5712
+ "owned_by": "mistralai"
5713
+ }
5714
+ },
5715
+ {
5716
+ "id": "open-mistral-7b",
5717
+ "name": "Open Mistral 7b",
5718
+ "provider": "mistral",
5719
+ "family": "mistral",
5720
+ "created_at": "2023-09-27 00:00:00 +0200",
5721
+ "context_window": 32768,
5722
+ "max_output_tokens": 8192,
5723
+ "knowledge_cutoff": null,
5724
+ "modalities": {
5725
+ "input": [
5726
+ "text"
5727
+ ],
5728
+ "output": [
5729
+ "text"
5730
+ ]
5731
+ },
5732
+ "capabilities": {
5733
+ "chat": true,
5734
+ "streaming": true,
5735
+ "tools": true,
5736
+ "vision": false,
5737
+ "json_mode": true
5738
+ },
5739
+ "pricing": {},
5740
+ "metadata": {
5741
+ "object": "model",
5742
+ "owned_by": "mistralai"
5743
+ }
5744
+ },
5745
+ {
5746
+ "id": "open-mistral-nemo",
5747
+ "name": "Open Mistral Nemo",
5748
+ "provider": "mistral",
5749
+ "family": "mistral",
5750
+ "created_at": "2024-07-18 00:00:00 +0200",
5751
+ "context_window": 32768,
5752
+ "max_output_tokens": 8192,
5753
+ "knowledge_cutoff": null,
5754
+ "modalities": {
5755
+ "input": [
5756
+ "text"
5757
+ ],
5758
+ "output": [
5759
+ "text"
5760
+ ]
5761
+ },
5762
+ "capabilities": {
5763
+ "chat": true,
5764
+ "streaming": true,
5765
+ "tools": true,
5766
+ "vision": false,
5767
+ "json_mode": true
5768
+ },
5769
+ "pricing": {},
5770
+ "metadata": {
5771
+ "object": "model",
5772
+ "owned_by": "mistralai"
5773
+ }
5774
+ },
5775
+ {
5776
+ "id": "open-mistral-nemo-2407",
5777
+ "name": "Open Mistral Nemo 2407",
5778
+ "provider": "mistral",
5779
+ "family": "mistral",
5780
+ "created_at": "2024-07-18 00:00:00 +0200",
5781
+ "context_window": 32768,
5782
+ "max_output_tokens": 8192,
5783
+ "knowledge_cutoff": null,
5784
+ "modalities": {
5785
+ "input": [
5786
+ "text"
5787
+ ],
5788
+ "output": [
5789
+ "text"
5790
+ ]
5791
+ },
5792
+ "capabilities": {
5793
+ "chat": true,
5794
+ "streaming": true,
5795
+ "tools": true,
5796
+ "vision": false,
5797
+ "json_mode": true
5798
+ },
5799
+ "pricing": {},
5800
+ "metadata": {
5801
+ "object": "model",
5802
+ "owned_by": "mistralai"
5803
+ }
5804
+ },
5805
+ {
5806
+ "id": "open-mixtral-8x22b",
5807
+ "name": "Open Mixtral 8x22b",
5808
+ "provider": "mistral",
5809
+ "family": "mistral",
5810
+ "created_at": "2024-04-17 00:00:00 +0200",
5811
+ "context_window": 32768,
5812
+ "max_output_tokens": 8192,
5813
+ "knowledge_cutoff": null,
5814
+ "modalities": {
5815
+ "input": [
5816
+ "text"
5817
+ ],
5818
+ "output": [
5819
+ "text"
5820
+ ]
5821
+ },
5822
+ "capabilities": {
5823
+ "chat": true,
5824
+ "streaming": true,
5825
+ "tools": true,
5826
+ "vision": false,
5827
+ "json_mode": true
5828
+ },
5829
+ "pricing": {},
5830
+ "metadata": {
5831
+ "object": "model",
5832
+ "owned_by": "mistralai"
5833
+ }
5834
+ },
5835
+ {
5836
+ "id": "open-mixtral-8x22b-2404",
5837
+ "name": "Open Mixtral 8x22b 2404",
5838
+ "provider": "mistral",
5839
+ "family": "mistral",
5840
+ "created_at": "2024-04-17 00:00:00 +0200",
5841
+ "context_window": 32768,
5842
+ "max_output_tokens": 8192,
5843
+ "knowledge_cutoff": null,
5844
+ "modalities": {
5845
+ "input": [
5846
+ "text"
5847
+ ],
5848
+ "output": [
5849
+ "text"
5850
+ ]
5851
+ },
5852
+ "capabilities": {
5853
+ "chat": true,
5854
+ "streaming": true,
5855
+ "tools": true,
5856
+ "vision": false,
5857
+ "json_mode": true
5858
+ },
5859
+ "pricing": {},
5860
+ "metadata": {
5861
+ "object": "model",
5862
+ "owned_by": "mistralai"
5863
+ }
5864
+ },
5865
+ {
5866
+ "id": "open-mixtral-8x7b",
5867
+ "name": "Open Mixtral 8x7b",
5868
+ "provider": "mistral",
5869
+ "family": "mistral",
5870
+ "created_at": "2023-12-11 00:00:00 +0100",
5871
+ "context_window": 32768,
5872
+ "max_output_tokens": 8192,
5873
+ "knowledge_cutoff": null,
5874
+ "modalities": {
5875
+ "input": [
5876
+ "text"
5877
+ ],
5878
+ "output": [
5879
+ "text"
5880
+ ]
5881
+ },
5882
+ "capabilities": {
5883
+ "chat": true,
5884
+ "streaming": true,
5885
+ "tools": true,
5886
+ "vision": false,
5887
+ "json_mode": true
5888
+ },
5889
+ "pricing": {},
5890
+ "metadata": {
5891
+ "object": "model",
5892
+ "owned_by": "mistralai"
4036
5893
  }
4037
5894
  },
4038
5895
  {
4039
- "id": "gemma-3n-e4b-it",
4040
- "name": "Gemma 3n E4B",
4041
- "provider": "gemini",
4042
- "family": "other",
4043
- "created_at": null,
4044
- "context_window": 8192,
4045
- "max_output_tokens": 2048,
5896
+ "id": "pixtral-12b",
5897
+ "name": "Pixtral 12B",
5898
+ "provider": "mistral",
5899
+ "family": "pixtral",
5900
+ "created_at": "2024-09-17 00:00:00 +0200",
5901
+ "context_window": 32768,
5902
+ "max_output_tokens": 8192,
4046
5903
  "knowledge_cutoff": null,
4047
5904
  "modalities": {
4048
5905
  "input": [
4049
- "text"
5906
+ "text",
5907
+ "image"
4050
5908
  ],
4051
5909
  "output": [
4052
5910
  "text"
4053
5911
  ]
4054
5912
  },
4055
- "capabilities": [
4056
- "streaming"
4057
- ],
4058
- "pricing": {
4059
- "text_tokens": {
4060
- "standard": {
4061
- "input_per_million": 0.075,
4062
- "output_per_million": 0.3
4063
- },
4064
- "batch": {
4065
- "input_per_million": 0.0375,
4066
- "output_per_million": 0.15
4067
- }
4068
- }
5913
+ "capabilities": {
5914
+ "chat": true,
5915
+ "streaming": true,
5916
+ "tools": true,
5917
+ "vision": true,
5918
+ "json_mode": true
4069
5919
  },
5920
+ "pricing": {},
4070
5921
  "metadata": {
4071
- "version": "001",
4072
- "description": null,
4073
- "supported_generation_methods": [
4074
- "generateContent",
4075
- "countTokens"
4076
- ]
5922
+ "object": "model",
5923
+ "owned_by": "mistralai"
4077
5924
  }
4078
5925
  },
4079
5926
  {
4080
- "id": "imagen-3.0-generate-002",
4081
- "name": "Imagen 3",
4082
- "provider": "gemini",
4083
- "family": "imagen-3.0-generate-002",
4084
- "created_at": null,
4085
- "context_window": null,
4086
- "max_output_tokens": null,
5927
+ "id": "pixtral-12b-2409",
5928
+ "name": "Pixtral 12B",
5929
+ "provider": "mistral",
5930
+ "family": "pixtral",
5931
+ "created_at": "2024-09-17 00:00:00 +0200",
5932
+ "context_window": 32768,
5933
+ "max_output_tokens": 8192,
4087
5934
  "knowledge_cutoff": null,
4088
5935
  "modalities": {
4089
5936
  "input": [
4090
- "text"
5937
+ "text",
5938
+ "image"
4091
5939
  ],
4092
5940
  "output": [
4093
- "image"
5941
+ "text"
4094
5942
  ]
4095
5943
  },
4096
- "capabilities": [],
4097
- "pricing": {
4098
- "text_tokens": {
4099
- "standard": {
4100
- "output_per_million": 0.03
4101
- }
4102
- }
5944
+ "capabilities": {
5945
+ "chat": true,
5946
+ "streaming": true,
5947
+ "tools": true,
5948
+ "vision": true,
5949
+ "json_mode": true
4103
5950
  },
5951
+ "pricing": {},
4104
5952
  "metadata": {
4105
- "version": "002",
4106
- "description": "Vertex served Imagen 3.0 002 model",
4107
- "supported_generation_methods": [
4108
- "predict"
4109
- ]
5953
+ "object": "model",
5954
+ "owned_by": "mistralai"
4110
5955
  }
4111
5956
  },
4112
5957
  {
4113
- "id": "imagen-4.0-generate-preview-06-06",
4114
- "name": "Imagen 4",
4115
- "provider": "gemini",
4116
- "family": "imagen-4.0-generate-preview-06-06",
4117
- "created_at": null,
4118
- "context_window": 480,
4119
- "max_output_tokens": null,
5958
+ "id": "pixtral-12b-latest",
5959
+ "name": "Pixtral 12B",
5960
+ "provider": "mistral",
5961
+ "family": "pixtral",
5962
+ "created_at": "2024-09-17 00:00:00 +0200",
5963
+ "context_window": 32768,
5964
+ "max_output_tokens": 8192,
4120
5965
  "knowledge_cutoff": null,
4121
5966
  "modalities": {
4122
5967
  "input": [
4123
- "text"
5968
+ "text",
5969
+ "image"
4124
5970
  ],
4125
5971
  "output": [
4126
- "image"
5972
+ "text"
4127
5973
  ]
4128
5974
  },
4129
- "capabilities": [],
4130
- "pricing": {
4131
- "text_tokens": {
4132
- "standard": {
4133
- "output_per_million": 0.04
4134
- }
4135
- }
5975
+ "capabilities": {
5976
+ "chat": true,
5977
+ "streaming": true,
5978
+ "tools": true,
5979
+ "vision": true,
5980
+ "json_mode": true
4136
5981
  },
5982
+ "pricing": {},
4137
5983
  "metadata": {
4138
- "version": "01",
4139
- "description": "Vertex served Imagen 4.0 model",
4140
- "supported_generation_methods": [
4141
- "predict"
4142
- ]
5984
+ "object": "model",
5985
+ "owned_by": "mistralai"
4143
5986
  }
4144
5987
  },
4145
5988
  {
4146
- "id": "imagen-4.0-ultra-generate-preview-06-06",
4147
- "name": "Imagen 4",
4148
- "provider": "gemini",
4149
- "family": "imagen-4.0-generate-preview-06-06",
4150
- "created_at": null,
4151
- "context_window": 480,
4152
- "max_output_tokens": null,
5989
+ "id": "pixtral-large-2411",
5990
+ "name": "Pixtral Large",
5991
+ "provider": "mistral",
5992
+ "family": "pixtral",
5993
+ "created_at": "2024-11-12 00:00:00 +0100",
5994
+ "context_window": 32768,
5995
+ "max_output_tokens": 8192,
4153
5996
  "knowledge_cutoff": null,
4154
5997
  "modalities": {
4155
5998
  "input": [
4156
- "text"
5999
+ "text",
6000
+ "image"
4157
6001
  ],
4158
6002
  "output": [
4159
- "image"
6003
+ "text"
4160
6004
  ]
4161
6005
  },
4162
- "capabilities": [],
4163
- "pricing": {
4164
- "text_tokens": {
4165
- "standard": {
4166
- "output_per_million": 0.04
4167
- }
4168
- }
6006
+ "capabilities": {
6007
+ "chat": true,
6008
+ "streaming": true,
6009
+ "tools": true,
6010
+ "vision": true,
6011
+ "json_mode": true
4169
6012
  },
6013
+ "pricing": {},
4170
6014
  "metadata": {
4171
- "version": "01",
4172
- "description": "Vertex served Imagen 4.0 ultra model",
4173
- "supported_generation_methods": [
4174
- "predict"
4175
- ]
6015
+ "object": "model",
6016
+ "owned_by": "mistralai"
4176
6017
  }
4177
6018
  },
4178
6019
  {
4179
- "id": "learnlm-2.0-flash-experimental",
4180
- "name": "LearnLM 2.0 Flash Experimental",
4181
- "provider": "gemini",
4182
- "family": "other",
4183
- "created_at": null,
4184
- "context_window": 1048576,
4185
- "max_output_tokens": 32768,
6020
+ "id": "pixtral-large-latest",
6021
+ "name": "Pixtral Large",
6022
+ "provider": "mistral",
6023
+ "family": "pixtral",
6024
+ "created_at": "2024-11-12 00:00:00 +0100",
6025
+ "context_window": 32768,
6026
+ "max_output_tokens": 8192,
4186
6027
  "knowledge_cutoff": null,
4187
6028
  "modalities": {
4188
6029
  "input": [
4189
6030
  "text",
4190
- "image",
4191
- "pdf"
6031
+ "image"
4192
6032
  ],
4193
6033
  "output": [
4194
6034
  "text"
4195
6035
  ]
4196
6036
  },
4197
- "capabilities": [
4198
- "streaming",
4199
- "function_calling",
4200
- "structured_output",
4201
- "batch",
4202
- "caching"
4203
- ],
4204
- "pricing": {
4205
- "text_tokens": {
4206
- "standard": {
4207
- "input_per_million": 0.075,
4208
- "output_per_million": 0.3
4209
- },
4210
- "batch": {
4211
- "input_per_million": 0.0375,
4212
- "output_per_million": 0.15
4213
- }
4214
- }
6037
+ "capabilities": {
6038
+ "chat": true,
6039
+ "streaming": true,
6040
+ "tools": true,
6041
+ "vision": true,
6042
+ "json_mode": true
4215
6043
  },
6044
+ "pricing": {},
4216
6045
  "metadata": {
4217
- "version": "2.0",
4218
- "description": "LearnLM 2.0 Flash Experimental",
4219
- "supported_generation_methods": [
4220
- "generateContent",
4221
- "countTokens"
4222
- ]
6046
+ "object": "model",
6047
+ "owned_by": "mistralai"
4223
6048
  }
4224
6049
  },
4225
6050
  {
4226
- "id": "models/text-embedding-004",
4227
- "name": "Legacy Embedding Models",
4228
- "provider": "gemini",
4229
- "family": "models/text-embedding-004",
6051
+ "id": "voxtral-mini-2507",
6052
+ "name": "Voxtral Mini 2507",
6053
+ "provider": "mistral",
6054
+ "family": "mistral",
4230
6055
  "created_at": null,
4231
- "context_window": 2048,
4232
- "max_output_tokens": null,
6056
+ "context_window": 32768,
6057
+ "max_output_tokens": 8192,
4233
6058
  "knowledge_cutoff": null,
4234
6059
  "modalities": {
4235
6060
  "input": [
4236
6061
  "text"
4237
6062
  ],
4238
6063
  "output": [
4239
- "embeddings"
6064
+ "text"
4240
6065
  ]
4241
6066
  },
4242
- "capabilities": [],
6067
+ "capabilities": {
6068
+ "chat": true,
6069
+ "streaming": true,
6070
+ "tools": true,
6071
+ "vision": false,
6072
+ "json_mode": true
6073
+ },
4243
6074
  "pricing": {},
4244
- "metadata": {}
6075
+ "metadata": {
6076
+ "object": "model",
6077
+ "owned_by": "mistralai"
6078
+ }
4245
6079
  },
4246
6080
  {
4247
- "id": "text-embedding-004",
4248
- "name": "Text Embedding 004",
4249
- "provider": "gemini",
4250
- "family": "embedding4",
6081
+ "id": "voxtral-mini-latest",
6082
+ "name": "Voxtral Mini Latest",
6083
+ "provider": "mistral",
6084
+ "family": "mistral",
4251
6085
  "created_at": null,
4252
- "context_window": 2048,
4253
- "max_output_tokens": 1,
6086
+ "context_window": 32768,
6087
+ "max_output_tokens": 8192,
4254
6088
  "knowledge_cutoff": null,
4255
6089
  "modalities": {
4256
6090
  "input": [
4257
6091
  "text"
4258
6092
  ],
4259
6093
  "output": [
4260
- "text",
4261
- "embeddings"
6094
+ "text"
4262
6095
  ]
4263
6096
  },
4264
- "capabilities": [
4265
- "streaming",
4266
- "batch"
4267
- ],
4268
- "pricing": {
4269
- "embeddings": {
4270
- "standard": {
4271
- "input_per_million": 0.002
4272
- }
4273
- }
6097
+ "capabilities": {
6098
+ "chat": true,
6099
+ "streaming": true,
6100
+ "tools": true,
6101
+ "vision": false,
6102
+ "json_mode": true
4274
6103
  },
6104
+ "pricing": {},
4275
6105
  "metadata": {
4276
- "version": "004",
4277
- "description": "Obtain a distributed representation of a text.",
4278
- "supported_generation_methods": [
4279
- "embedContent"
4280
- ]
6106
+ "object": "model",
6107
+ "owned_by": "mistralai"
4281
6108
  }
4282
6109
  },
4283
6110
  {
4284
- "id": "veo-2.0-generate-001",
4285
- "name": "Veo 2",
4286
- "provider": "gemini",
4287
- "family": "veo-2.0-generate-001",
6111
+ "id": "voxtral-small-2507",
6112
+ "name": "Voxtral Small 2507",
6113
+ "provider": "mistral",
6114
+ "family": "mistral",
4288
6115
  "created_at": null,
4289
- "context_window": null,
4290
- "max_output_tokens": null,
6116
+ "context_window": 32768,
6117
+ "max_output_tokens": 8192,
4291
6118
  "knowledge_cutoff": null,
4292
6119
  "modalities": {
4293
6120
  "input": [
4294
- "image",
4295
6121
  "text"
4296
6122
  ],
4297
- "output": []
6123
+ "output": [
6124
+ "text"
6125
+ ]
4298
6126
  },
4299
- "capabilities": [],
4300
- "pricing": {
4301
- "text_tokens": {
4302
- "standard": {
4303
- "output_per_million": 0.35
4304
- }
4305
- }
6127
+ "capabilities": {
6128
+ "chat": true,
6129
+ "streaming": true,
6130
+ "tools": true,
6131
+ "vision": false,
6132
+ "json_mode": true
4306
6133
  },
6134
+ "pricing": {},
4307
6135
  "metadata": {
4308
- "version": "2.0",
4309
- "description": "Vertex served Veo 2 model. Access to this model requires billing to be enabled on the associated Google Cloud Platform account. Please visit https://console.cloud.google.com/billing to enable it.",
4310
- "supported_generation_methods": [
4311
- "predictLongRunning"
4312
- ]
6136
+ "object": "model",
6137
+ "owned_by": "mistralai"
4313
6138
  }
4314
6139
  },
4315
6140
  {
4316
- "id": "veo-3.0-generate-preview",
4317
- "name": "Veo 3 Preview",
4318
- "provider": "gemini",
4319
- "family": "veo-3.0-generate-preview",
6141
+ "id": "voxtral-small-latest",
6142
+ "name": "Voxtral Small Latest",
6143
+ "provider": "mistral",
6144
+ "family": "mistral",
4320
6145
  "created_at": null,
4321
- "context_window": 1024,
4322
- "max_output_tokens": null,
6146
+ "context_window": 32768,
6147
+ "max_output_tokens": 8192,
4323
6148
  "knowledge_cutoff": null,
4324
6149
  "modalities": {
4325
6150
  "input": [
4326
6151
  "text"
4327
6152
  ],
4328
6153
  "output": [
4329
- "audio"
6154
+ "text"
4330
6155
  ]
4331
6156
  },
4332
- "capabilities": [],
4333
- "pricing": {
4334
- "text_tokens": {
4335
- "standard": {
4336
- "output_per_million": 0.75
4337
- }
4338
- }
6157
+ "capabilities": {
6158
+ "chat": true,
6159
+ "streaming": true,
6160
+ "tools": true,
6161
+ "vision": false,
6162
+ "json_mode": true
4339
6163
  },
4340
- "metadata": {}
6164
+ "pricing": {},
6165
+ "metadata": {
6166
+ "object": "model",
6167
+ "owned_by": "mistralai"
6168
+ }
4341
6169
  },
4342
6170
  {
4343
6171
  "id": "babbage-002",
@@ -7195,7 +9023,8 @@
7195
9023
  ],
7196
9024
  "output": [
7197
9025
  "audio",
7198
- "embeddings"
9026
+ "embeddings",
9027
+ "text"
7199
9028
  ]
7200
9029
  },
7201
9030
  "capabilities": [],
@@ -7268,7 +9097,7 @@
7268
9097
  "pricing": {
7269
9098
  "text_tokens": {
7270
9099
  "standard": {
7271
- "input_per_million": 30.0
9100
+ "output_per_million": 30.0
7272
9101
  }
7273
9102
  }
7274
9103
  },
@@ -17625,10 +19454,7 @@
17625
19454
  "presence_penalty",
17626
19455
  "response_format",
17627
19456
  "structured_outputs",
17628
- "seed",
17629
- "top_logprobs",
17630
- "logprobs",
17631
- "logit_bias"
19457
+ "seed"
17632
19458
  ]
17633
19459
  }
17634
19460
  },
@@ -18557,16 +20383,16 @@
18557
20383
  "top_p",
18558
20384
  "tools",
18559
20385
  "tool_choice",
20386
+ "structured_outputs",
20387
+ "response_format",
18560
20388
  "stop",
18561
20389
  "frequency_penalty",
18562
20390
  "presence_penalty",
18563
- "repetition_penalty",
18564
- "response_format",
18565
20391
  "top_k",
18566
- "seed",
18567
- "min_p",
18568
- "structured_outputs",
20392
+ "repetition_penalty",
18569
20393
  "logit_bias",
20394
+ "min_p",
20395
+ "seed",
18570
20396
  "logprobs",
18571
20397
  "top_logprobs"
18572
20398
  ]
@@ -19048,8 +20874,8 @@
19048
20874
  "provider": "openrouter",
19049
20875
  "family": "neversleep",
19050
20876
  "created_at": "2023-11-26 01:00:00 +0100",
19051
- "context_window": 8192,
19052
- "max_output_tokens": 2048,
20877
+ "context_window": 4096,
20878
+ "max_output_tokens": null,
19053
20879
  "knowledge_cutoff": null,
19054
20880
  "modalities": {
19055
20881
  "input": [
@@ -19061,13 +20887,14 @@
19061
20887
  },
19062
20888
  "capabilities": [
19063
20889
  "streaming",
20890
+ "structured_output",
19064
20891
  "predicted_outputs"
19065
20892
  ],
19066
20893
  "pricing": {
19067
20894
  "text_tokens": {
19068
20895
  "standard": {
19069
- "input_per_million": 1.25,
19070
- "output_per_million": 2.0
20896
+ "input_per_million": 1.0,
20897
+ "output_per_million": 1.75
19071
20898
  }
19072
20899
  }
19073
20900
  },
@@ -19085,8 +20912,8 @@
19085
20912
  "instruct_type": "alpaca"
19086
20913
  },
19087
20914
  "top_provider": {
19088
- "context_length": 8192,
19089
- "max_completion_tokens": 2048,
20915
+ "context_length": 4096,
20916
+ "max_completion_tokens": null,
19090
20917
  "is_moderated": false
19091
20918
  },
19092
20919
  "per_request_limits": null,
@@ -19097,6 +20924,8 @@
19097
20924
  "stop",
19098
20925
  "frequency_penalty",
19099
20926
  "presence_penalty",
20927
+ "response_format",
20928
+ "structured_outputs",
19100
20929
  "repetition_penalty",
19101
20930
  "logit_bias",
19102
20931
  "top_k",
@@ -19348,11 +21177,12 @@
19348
21177
  "stop",
19349
21178
  "frequency_penalty",
19350
21179
  "presence_penalty",
21180
+ "response_format",
21181
+ "structured_outputs",
19351
21182
  "seed",
19352
21183
  "logit_bias",
19353
21184
  "logprobs",
19354
21185
  "top_logprobs",
19355
- "response_format",
19356
21186
  "min_p",
19357
21187
  "repetition_penalty",
19358
21188
  "top_k"
@@ -21551,114 +23381,6 @@
21551
23381
  ]
21552
23382
  }
21553
23383
  },
21554
- {
21555
- "id": "openai/o1-preview",
21556
- "name": "OpenAI: o1-preview",
21557
- "provider": "openrouter",
21558
- "family": "openai",
21559
- "created_at": "2024-09-12 02:00:00 +0200",
21560
- "context_window": 128000,
21561
- "max_output_tokens": 32768,
21562
- "knowledge_cutoff": null,
21563
- "modalities": {
21564
- "input": [
21565
- "text"
21566
- ],
21567
- "output": [
21568
- "text"
21569
- ]
21570
- },
21571
- "capabilities": [
21572
- "streaming"
21573
- ],
21574
- "pricing": {
21575
- "text_tokens": {
21576
- "standard": {
21577
- "input_per_million": 15.0,
21578
- "output_per_million": 60.0,
21579
- "cached_input_per_million": 7.5
21580
- }
21581
- }
21582
- },
21583
- "metadata": {
21584
- "description": "The latest and strongest model family from OpenAI, o1 is designed to spend more time thinking before responding.\n\nThe o1 models are optimized for math, science, programming, and other STEM-related tasks. They consistently exhibit PhD-level accuracy on benchmarks in physics, chemistry, and biology. Learn more in the [launch announcement](https://openai.com/o1).\n\nNote: This model is currently experimental and not suitable for production use-cases, and may be heavily rate-limited.",
21585
- "architecture": {
21586
- "modality": "text->text",
21587
- "input_modalities": [
21588
- "text"
21589
- ],
21590
- "output_modalities": [
21591
- "text"
21592
- ],
21593
- "tokenizer": "GPT",
21594
- "instruct_type": null
21595
- },
21596
- "top_provider": {
21597
- "context_length": 128000,
21598
- "max_completion_tokens": 32768,
21599
- "is_moderated": true
21600
- },
21601
- "per_request_limits": null,
21602
- "supported_parameters": [
21603
- "seed",
21604
- "max_tokens"
21605
- ]
21606
- }
21607
- },
21608
- {
21609
- "id": "openai/o1-preview-2024-09-12",
21610
- "name": "OpenAI: o1-preview (2024-09-12)",
21611
- "provider": "openrouter",
21612
- "family": "openai",
21613
- "created_at": "2024-09-12 02:00:00 +0200",
21614
- "context_window": 128000,
21615
- "max_output_tokens": 32768,
21616
- "knowledge_cutoff": null,
21617
- "modalities": {
21618
- "input": [
21619
- "text"
21620
- ],
21621
- "output": [
21622
- "text"
21623
- ]
21624
- },
21625
- "capabilities": [
21626
- "streaming"
21627
- ],
21628
- "pricing": {
21629
- "text_tokens": {
21630
- "standard": {
21631
- "input_per_million": 15.0,
21632
- "output_per_million": 60.0,
21633
- "cached_input_per_million": 7.5
21634
- }
21635
- }
21636
- },
21637
- "metadata": {
21638
- "description": "The latest and strongest model family from OpenAI, o1 is designed to spend more time thinking before responding.\n\nThe o1 models are optimized for math, science, programming, and other STEM-related tasks. They consistently exhibit PhD-level accuracy on benchmarks in physics, chemistry, and biology. Learn more in the [launch announcement](https://openai.com/o1).\n\nNote: This model is currently experimental and not suitable for production use-cases, and may be heavily rate-limited.",
21639
- "architecture": {
21640
- "modality": "text->text",
21641
- "input_modalities": [
21642
- "text"
21643
- ],
21644
- "output_modalities": [
21645
- "text"
21646
- ],
21647
- "tokenizer": "GPT",
21648
- "instruct_type": null
21649
- },
21650
- "top_provider": {
21651
- "context_length": 128000,
21652
- "max_completion_tokens": 32768,
21653
- "is_moderated": true
21654
- },
21655
- "per_request_limits": null,
21656
- "supported_parameters": [
21657
- "seed",
21658
- "max_tokens"
21659
- ]
21660
- }
21661
- },
21662
23384
  {
21663
23385
  "id": "openai/o1-pro",
21664
23386
  "name": "OpenAI: o1-pro",
@@ -22189,6 +23911,59 @@
22189
23911
  "supported_parameters": []
22190
23912
  }
22191
23913
  },
23914
+ {
23915
+ "id": "openrouter/horizon-alpha",
23916
+ "name": "Horizon Alpha",
23917
+ "provider": "openrouter",
23918
+ "family": "openrouter",
23919
+ "created_at": "2025-07-31 00:18:04 +0200",
23920
+ "context_window": 256000,
23921
+ "max_output_tokens": 128000,
23922
+ "knowledge_cutoff": null,
23923
+ "modalities": {
23924
+ "input": [
23925
+ "image",
23926
+ "text"
23927
+ ],
23928
+ "output": [
23929
+ "text"
23930
+ ]
23931
+ },
23932
+ "capabilities": [
23933
+ "streaming",
23934
+ "function_calling",
23935
+ "structured_output"
23936
+ ],
23937
+ "pricing": {},
23938
+ "metadata": {
23939
+ "description": "This is a cloaked model provided to the community to gather feedback.\n\nNote: It’s free to use during this testing period, and prompts and completions are logged by the model creator for feedback and training.",
23940
+ "architecture": {
23941
+ "modality": "text+image->text",
23942
+ "input_modalities": [
23943
+ "image",
23944
+ "text"
23945
+ ],
23946
+ "output_modalities": [
23947
+ "text"
23948
+ ],
23949
+ "tokenizer": "Other",
23950
+ "instruct_type": null
23951
+ },
23952
+ "top_provider": {
23953
+ "context_length": 256000,
23954
+ "max_completion_tokens": 128000,
23955
+ "is_moderated": false
23956
+ },
23957
+ "per_request_limits": null,
23958
+ "supported_parameters": [
23959
+ "max_tokens",
23960
+ "tool_choice",
23961
+ "tools",
23962
+ "response_format",
23963
+ "structured_outputs"
23964
+ ]
23965
+ }
23966
+ },
22192
23967
  {
22193
23968
  "id": "perplexity/r1-1776",
22194
23969
  "name": "Perplexity: R1 1776",
@@ -23769,83 +25544,10 @@
23769
25544
  "id": "qwen/qwen3-235b-a22b",
23770
25545
  "name": "Qwen: Qwen3 235B A22B",
23771
25546
  "provider": "openrouter",
23772
- "family": "qwen",
23773
- "created_at": "2025-04-28 23:29:17 +0200",
23774
- "context_window": 40960,
23775
- "max_output_tokens": 40960,
23776
- "knowledge_cutoff": null,
23777
- "modalities": {
23778
- "input": [
23779
- "text"
23780
- ],
23781
- "output": [
23782
- "text"
23783
- ]
23784
- },
23785
- "capabilities": [
23786
- "streaming",
23787
- "function_calling",
23788
- "structured_output",
23789
- "predicted_outputs"
23790
- ],
23791
- "pricing": {
23792
- "text_tokens": {
23793
- "standard": {
23794
- "input_per_million": 0.13,
23795
- "output_per_million": 0.6
23796
- }
23797
- }
23798
- },
23799
- "metadata": {
23800
- "description": "Qwen3-235B-A22B is a 235B parameter mixture-of-experts (MoE) model developed by Qwen, activating 22B parameters per forward pass. It supports seamless switching between a \"thinking\" mode for complex reasoning, math, and code tasks, and a \"non-thinking\" mode for general conversational efficiency. The model demonstrates strong reasoning ability, multilingual support (100+ languages and dialects), advanced instruction-following, and agent tool-calling capabilities. It natively handles a 32K token context window and extends up to 131K tokens using YaRN-based scaling.",
23801
- "architecture": {
23802
- "modality": "text->text",
23803
- "input_modalities": [
23804
- "text"
23805
- ],
23806
- "output_modalities": [
23807
- "text"
23808
- ],
23809
- "tokenizer": "Qwen3",
23810
- "instruct_type": "qwen3"
23811
- },
23812
- "top_provider": {
23813
- "context_length": 40960,
23814
- "max_completion_tokens": 40960,
23815
- "is_moderated": false
23816
- },
23817
- "per_request_limits": null,
23818
- "supported_parameters": [
23819
- "max_tokens",
23820
- "temperature",
23821
- "top_p",
23822
- "tools",
23823
- "tool_choice",
23824
- "reasoning",
23825
- "include_reasoning",
23826
- "structured_outputs",
23827
- "response_format",
23828
- "stop",
23829
- "frequency_penalty",
23830
- "presence_penalty",
23831
- "seed",
23832
- "top_k",
23833
- "min_p",
23834
- "repetition_penalty",
23835
- "logprobs",
23836
- "top_logprobs",
23837
- "logit_bias"
23838
- ]
23839
- }
23840
- },
23841
- {
23842
- "id": "qwen/qwen3-235b-a22b-2507",
23843
- "name": "Qwen: Qwen3 235B A22B Instruct 2507",
23844
- "provider": "openrouter",
23845
- "family": "qwen",
23846
- "created_at": "2025-07-21 19:39:15 +0200",
23847
- "context_window": 262144,
23848
- "max_output_tokens": 262144,
25547
+ "family": "qwen",
25548
+ "created_at": "2025-04-28 23:29:17 +0200",
25549
+ "context_window": 40960,
25550
+ "max_output_tokens": 40960,
23849
25551
  "knowledge_cutoff": null,
23850
25552
  "modalities": {
23851
25553
  "input": [
@@ -23864,13 +25566,13 @@
23864
25566
  "pricing": {
23865
25567
  "text_tokens": {
23866
25568
  "standard": {
23867
- "input_per_million": 0.12,
23868
- "output_per_million": 0.59
25569
+ "input_per_million": 0.13,
25570
+ "output_per_million": 0.6
23869
25571
  }
23870
25572
  }
23871
25573
  },
23872
25574
  "metadata": {
23873
- "description": "Qwen3-235B-A22B-Instruct-2507 is a multilingual, instruction-tuned mixture-of-experts language model based on the Qwen3-235B architecture, with 22B active parameters per forward pass. It is optimized for general-purpose text generation, including instruction following, logical reasoning, math, code, and tool usage. The model supports a native 262K context length and does not implement \"thinking mode\" (<think> blocks).\n\nCompared to its base variant, this version delivers significant gains in knowledge coverage, long-context reasoning, coding benchmarks, and alignment with open-ended tasks. It is particularly strong on multilingual understanding, math reasoning (e.g., AIME, HMMT), and alignment evaluations like Arena-Hard and WritingBench.",
25575
+ "description": "Qwen3-235B-A22B is a 235B parameter mixture-of-experts (MoE) model developed by Qwen, activating 22B parameters per forward pass. It supports seamless switching between a \"thinking\" mode for complex reasoning, math, and code tasks, and a \"non-thinking\" mode for general conversational efficiency. The model demonstrates strong reasoning ability, multilingual support (100+ languages and dialects), advanced instruction-following, and agent tool-calling capabilities. It natively handles a 32K token context window and extends up to 131K tokens using YaRN-based scaling.",
23874
25576
  "architecture": {
23875
25577
  "modality": "text->text",
23876
25578
  "input_modalities": [
@@ -23880,11 +25582,11 @@
23880
25582
  "text"
23881
25583
  ],
23882
25584
  "tokenizer": "Qwen3",
23883
- "instruct_type": null
25585
+ "instruct_type": "qwen3"
23884
25586
  },
23885
25587
  "top_provider": {
23886
- "context_length": 262144,
23887
- "max_completion_tokens": 262144,
25588
+ "context_length": 40960,
25589
+ "max_completion_tokens": 40960,
23888
25590
  "is_moderated": false
23889
25591
  },
23890
25592
  "per_request_limits": null,
@@ -23892,17 +25594,19 @@
23892
25594
  "max_tokens",
23893
25595
  "temperature",
23894
25596
  "top_p",
23895
- "seed",
23896
25597
  "tools",
23897
25598
  "tool_choice",
25599
+ "reasoning",
25600
+ "include_reasoning",
25601
+ "structured_outputs",
25602
+ "response_format",
23898
25603
  "stop",
23899
25604
  "frequency_penalty",
23900
25605
  "presence_penalty",
23901
- "repetition_penalty",
23902
- "response_format",
25606
+ "seed",
23903
25607
  "top_k",
23904
25608
  "min_p",
23905
- "structured_outputs",
25609
+ "repetition_penalty",
23906
25610
  "logit_bias",
23907
25611
  "logprobs",
23908
25612
  "top_logprobs"
@@ -23910,8 +25614,8 @@
23910
25614
  }
23911
25615
  },
23912
25616
  {
23913
- "id": "qwen/qwen3-235b-a22b-2507:free",
23914
- "name": "Qwen: Qwen3 235B A22B Instruct 2507 (free)",
25617
+ "id": "qwen/qwen3-235b-a22b-2507",
25618
+ "name": "Qwen: Qwen3 235B A22B Instruct 2507",
23915
25619
  "provider": "openrouter",
23916
25620
  "family": "qwen",
23917
25621
  "created_at": "2025-07-21 19:39:15 +0200",
@@ -23928,9 +25632,18 @@
23928
25632
  },
23929
25633
  "capabilities": [
23930
25634
  "streaming",
25635
+ "function_calling",
25636
+ "structured_output",
23931
25637
  "predicted_outputs"
23932
25638
  ],
23933
- "pricing": {},
25639
+ "pricing": {
25640
+ "text_tokens": {
25641
+ "standard": {
25642
+ "input_per_million": 0.1179,
25643
+ "output_per_million": 0.1179
25644
+ }
25645
+ }
25646
+ },
23934
25647
  "metadata": {
23935
25648
  "description": "Qwen3-235B-A22B-Instruct-2507 is a multilingual, instruction-tuned mixture-of-experts language model based on the Qwen3-235B architecture, with 22B active parameters per forward pass. It is optimized for general-purpose text generation, including instruction following, logical reasoning, math, code, and tool usage. The model supports a native 262K context length and does not implement \"thinking mode\" (<think> blocks).\n\nCompared to its base variant, this version delivers significant gains in knowledge coverage, long-context reasoning, coding benchmarks, and alignment with open-ended tasks. It is particularly strong on multilingual understanding, math reasoning (e.g., AIME, HMMT), and alignment evaluations like Arena-Hard and WritingBench.",
23936
25649
  "architecture": {
@@ -23954,15 +25667,19 @@
23954
25667
  "max_tokens",
23955
25668
  "temperature",
23956
25669
  "top_p",
25670
+ "seed",
25671
+ "tools",
25672
+ "tool_choice",
23957
25673
  "stop",
23958
25674
  "frequency_penalty",
23959
25675
  "presence_penalty",
23960
- "seed",
25676
+ "repetition_penalty",
25677
+ "response_format",
23961
25678
  "top_k",
23962
25679
  "min_p",
23963
- "repetition_penalty",
23964
- "logprobs",
25680
+ "structured_outputs",
23965
25681
  "logit_bias",
25682
+ "logprobs",
23966
25683
  "top_logprobs"
23967
25684
  ]
23968
25685
  }
@@ -24178,6 +25895,64 @@
24178
25895
  ]
24179
25896
  }
24180
25897
  },
25898
+ {
25899
+ "id": "qwen/qwen3-30b-a3b-instruct-2507",
25900
+ "name": "Qwen: Qwen3 30B A3B Instruct 2507",
25901
+ "provider": "openrouter",
25902
+ "family": "qwen",
25903
+ "created_at": "2025-07-29 18:36:05 +0200",
25904
+ "context_window": 131072,
25905
+ "max_output_tokens": 32768,
25906
+ "knowledge_cutoff": null,
25907
+ "modalities": {
25908
+ "input": [
25909
+ "text"
25910
+ ],
25911
+ "output": [
25912
+ "text"
25913
+ ]
25914
+ },
25915
+ "capabilities": [
25916
+ "streaming",
25917
+ "structured_output"
25918
+ ],
25919
+ "pricing": {
25920
+ "text_tokens": {
25921
+ "standard": {
25922
+ "input_per_million": 0.19999999999999998,
25923
+ "output_per_million": 0.7999999999999999
25924
+ }
25925
+ }
25926
+ },
25927
+ "metadata": {
25928
+ "description": "Qwen3-30B-A3B-Instruct-2507 is a 30.5B-parameter mixture-of-experts language model from Qwen, with 3.3B active parameters per inference. It operates in non-thinking mode and is designed for high-quality instruction following, multilingual understanding, and agentic tool use. Post-trained on instruction data, it demonstrates competitive performance across reasoning (AIME, ZebraLogic), coding (MultiPL-E, LiveCodeBench), and alignment (IFEval, WritingBench) benchmarks. It outperforms its non-instruct variant on subjective and open-ended tasks while retaining strong factual and coding performance.",
25929
+ "architecture": {
25930
+ "modality": "text->text",
25931
+ "input_modalities": [
25932
+ "text"
25933
+ ],
25934
+ "output_modalities": [
25935
+ "text"
25936
+ ],
25937
+ "tokenizer": "Qwen3",
25938
+ "instruct_type": null
25939
+ },
25940
+ "top_provider": {
25941
+ "context_length": 131072,
25942
+ "max_completion_tokens": 32768,
25943
+ "is_moderated": false
25944
+ },
25945
+ "per_request_limits": null,
25946
+ "supported_parameters": [
25947
+ "max_tokens",
25948
+ "temperature",
25949
+ "top_p",
25950
+ "seed",
25951
+ "response_format",
25952
+ "presence_penalty"
25953
+ ]
25954
+ }
25955
+ },
24181
25956
  {
24182
25957
  "id": "qwen/qwen3-30b-a3b:free",
24183
25958
  "name": "Qwen: Qwen3 30B A3B (free)",
@@ -25892,8 +27667,8 @@
25892
27667
  "provider": "openrouter",
25893
27668
  "family": "thedrummer",
25894
27669
  "created_at": "2025-06-29 16:45:47 +0200",
25895
- "context_window": 131072,
25896
- "max_output_tokens": 131072,
27670
+ "context_window": 16384,
27671
+ "max_output_tokens": null,
25897
27672
  "knowledge_cutoff": null,
25898
27673
  "modalities": {
25899
27674
  "input": [
@@ -25904,13 +27679,14 @@
25904
27679
  ]
25905
27680
  },
25906
27681
  "capabilities": [
25907
- "streaming"
27682
+ "streaming",
27683
+ "structured_output"
25908
27684
  ],
25909
27685
  "pricing": {
25910
27686
  "text_tokens": {
25911
27687
  "standard": {
25912
- "input_per_million": 0.5,
25913
- "output_per_million": 0.7999999999999999
27688
+ "input_per_million": 0.39999999999999997,
27689
+ "output_per_million": 0.7
25914
27690
  }
25915
27691
  }
25916
27692
  },
@@ -25928,8 +27704,8 @@
25928
27704
  "instruct_type": null
25929
27705
  },
25930
27706
  "top_provider": {
25931
- "context_length": 131072,
25932
- "max_completion_tokens": 131072,
27707
+ "context_length": 16384,
27708
+ "max_completion_tokens": null,
25933
27709
  "is_moderated": false
25934
27710
  },
25935
27711
  "per_request_limits": null,
@@ -25943,7 +27719,9 @@
25943
27719
  "repetition_penalty",
25944
27720
  "seed",
25945
27721
  "stop",
25946
- "top_k"
27722
+ "top_k",
27723
+ "response_format",
27724
+ "structured_outputs"
25947
27725
  ]
25948
27726
  }
25949
27727
  },
@@ -26084,7 +27862,7 @@
26084
27862
  "provider": "openrouter",
26085
27863
  "family": "thedrummer",
26086
27864
  "created_at": "2025-03-10 20:56:06 +0100",
26087
- "context_window": 16384,
27865
+ "context_window": 32768,
26088
27866
  "max_output_tokens": null,
26089
27867
  "knowledge_cutoff": null,
26090
27868
  "modalities": {
@@ -26103,8 +27881,8 @@
26103
27881
  "pricing": {
26104
27882
  "text_tokens": {
26105
27883
  "standard": {
26106
- "input_per_million": 0.01512,
26107
- "output_per_million": 0.01512
27884
+ "input_per_million": 0.0728,
27885
+ "output_per_million": 0.0728
26108
27886
  }
26109
27887
  }
26110
27888
  },
@@ -26122,7 +27900,7 @@
26122
27900
  "instruct_type": null
26123
27901
  },
26124
27902
  "top_provider": {
26125
- "context_length": 16384,
27903
+ "context_length": 32768,
26126
27904
  "max_completion_tokens": null,
26127
27905
  "is_moderated": false
26128
27906
  },
@@ -27551,14 +29329,14 @@
27551
29329
  "tool_choice",
27552
29330
  "reasoning",
27553
29331
  "include_reasoning",
29332
+ "stop",
27554
29333
  "frequency_penalty",
27555
- "min_p",
27556
29334
  "presence_penalty",
27557
29335
  "repetition_penalty",
27558
- "seed",
27559
- "stop",
27560
- "top_k",
27561
29336
  "response_format",
29337
+ "top_k",
29338
+ "seed",
29339
+ "min_p",
27562
29340
  "logprobs",
27563
29341
  "logit_bias",
27564
29342
  "top_logprobs"
@@ -27693,5 +29471,168 @@
27693
29471
  "top_logprobs"
27694
29472
  ]
27695
29473
  }
29474
+ },
29475
+ {
29476
+ "id": "sonar",
29477
+ "name": "Sonar",
29478
+ "provider": "perplexity",
29479
+ "family": "sonar",
29480
+ "created_at": "2025-08-01 00:17:19 +0200",
29481
+ "context_window": 128000,
29482
+ "max_output_tokens": 4096,
29483
+ "knowledge_cutoff": null,
29484
+ "modalities": {
29485
+ "input": [
29486
+ "text"
29487
+ ],
29488
+ "output": [
29489
+ "text"
29490
+ ]
29491
+ },
29492
+ "capabilities": [
29493
+ "streaming",
29494
+ "json_mode",
29495
+ "vision"
29496
+ ],
29497
+ "pricing": {
29498
+ "text_tokens": {
29499
+ "standard": {
29500
+ "input_per_million": 1.0,
29501
+ "output_per_million": 1.0
29502
+ }
29503
+ }
29504
+ },
29505
+ "metadata": {}
29506
+ },
29507
+ {
29508
+ "id": "sonar-deep-research",
29509
+ "name": "Sonar Deep Research",
29510
+ "provider": "perplexity",
29511
+ "family": "sonar_deep_research",
29512
+ "created_at": "2025-08-01 00:17:19 +0200",
29513
+ "context_window": 128000,
29514
+ "max_output_tokens": 4096,
29515
+ "knowledge_cutoff": null,
29516
+ "modalities": {
29517
+ "input": [
29518
+ "text"
29519
+ ],
29520
+ "output": [
29521
+ "text"
29522
+ ]
29523
+ },
29524
+ "capabilities": [
29525
+ "streaming",
29526
+ "json_mode",
29527
+ "vision"
29528
+ ],
29529
+ "pricing": {
29530
+ "text_tokens": {
29531
+ "standard": {
29532
+ "input_per_million": 2.0,
29533
+ "output_per_million": 8.0,
29534
+ "citation_per_million": 2.0,
29535
+ "reasoning_per_million": 3.0,
29536
+ "search_per_thousand": 5.0
29537
+ }
29538
+ }
29539
+ },
29540
+ "metadata": {}
29541
+ },
29542
+ {
29543
+ "id": "sonar-pro",
29544
+ "name": "Sonar Pro",
29545
+ "provider": "perplexity",
29546
+ "family": "sonar_pro",
29547
+ "created_at": "2025-08-01 00:17:19 +0200",
29548
+ "context_window": 200000,
29549
+ "max_output_tokens": 8192,
29550
+ "knowledge_cutoff": null,
29551
+ "modalities": {
29552
+ "input": [
29553
+ "text"
29554
+ ],
29555
+ "output": [
29556
+ "text"
29557
+ ]
29558
+ },
29559
+ "capabilities": [
29560
+ "streaming",
29561
+ "json_mode",
29562
+ "vision"
29563
+ ],
29564
+ "pricing": {
29565
+ "text_tokens": {
29566
+ "standard": {
29567
+ "input_per_million": 3.0,
29568
+ "output_per_million": 15.0
29569
+ }
29570
+ }
29571
+ },
29572
+ "metadata": {}
29573
+ },
29574
+ {
29575
+ "id": "sonar-reasoning",
29576
+ "name": "Sonar Reasoning",
29577
+ "provider": "perplexity",
29578
+ "family": "sonar_reasoning",
29579
+ "created_at": "2025-08-01 00:17:19 +0200",
29580
+ "context_window": 128000,
29581
+ "max_output_tokens": 4096,
29582
+ "knowledge_cutoff": null,
29583
+ "modalities": {
29584
+ "input": [
29585
+ "text"
29586
+ ],
29587
+ "output": [
29588
+ "text"
29589
+ ]
29590
+ },
29591
+ "capabilities": [
29592
+ "streaming",
29593
+ "json_mode",
29594
+ "vision"
29595
+ ],
29596
+ "pricing": {
29597
+ "text_tokens": {
29598
+ "standard": {
29599
+ "input_per_million": 1.0,
29600
+ "output_per_million": 5.0
29601
+ }
29602
+ }
29603
+ },
29604
+ "metadata": {}
29605
+ },
29606
+ {
29607
+ "id": "sonar-reasoning-pro",
29608
+ "name": "Sonar Reasoning Pro",
29609
+ "provider": "perplexity",
29610
+ "family": "sonar_reasoning_pro",
29611
+ "created_at": "2025-08-01 00:17:19 +0200",
29612
+ "context_window": 128000,
29613
+ "max_output_tokens": 8192,
29614
+ "knowledge_cutoff": null,
29615
+ "modalities": {
29616
+ "input": [
29617
+ "text"
29618
+ ],
29619
+ "output": [
29620
+ "text"
29621
+ ]
29622
+ },
29623
+ "capabilities": [
29624
+ "streaming",
29625
+ "json_mode",
29626
+ "vision"
29627
+ ],
29628
+ "pricing": {
29629
+ "text_tokens": {
29630
+ "standard": {
29631
+ "input_per_million": 2.0,
29632
+ "output_per_million": 8.0
29633
+ }
29634
+ }
29635
+ },
29636
+ "metadata": {}
27696
29637
  }
27697
29638
  ]