pi-fireworks-provider 1.5.2 → 1.5.4

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.
Files changed (5) hide show
  1. package/README.md +15 -13
  2. package/custom-models.json +35 -6
  3. package/models.json +16 -0
  4. package/package.json +12 -11
  5. package/patch.json +283 -52
package/README.md CHANGED
@@ -15,7 +15,7 @@ _Kimi, MiniMax, GLM, DeepSeek, GPT-OSS — via Fireworks AI's Anthropic Messages
15
15
 
16
16
  ## Features
17
17
 
18
- - **41+ AI Models** including Kimi K2.5, MiniMax M2.5, GLM 4.5/4.7/5, DeepSeek V3.1/V3.2, DeepSeek V4 Flash, and GPT-OSS
18
+ - **43+ AI Models** including Kimi K2.5, MiniMax M2.5, GLM 4.5/4.7/5, DeepSeek V3.1/V3.2, DeepSeek V4 Flash, and GPT-OSS
19
19
  - **Dual API support** via Fireworks AI's Anthropic Messages and OpenAI-compatible completions endpoints (per-model routing, matching pi core's Fireworks provider)
20
20
  - **Service tiers** — toggle Fireworks `priority` vs `standard` per request on supported models (with priority pricing reflected in cost tracking), via a keybinding, `/fireworks-tier`, and a footer status area
21
21
  - **Preserved thinking** — toggle Fireworks' `reasoning_history: "preserved"` so prior assistant reasoning is retained across turns (better multi-turn recall; uses more tokens), via the `/fireworks-settings` panel, with a model-select notification. Matches neuralwatt/makora's settings-only UX, adapted to Fireworks' single global `reasoning_history` knob
@@ -75,10 +75,11 @@ pi
75
75
  | DeepSeek V3.1 | Text | 164K | 164K | $0.56 | $1.68 |
76
76
  | DeepSeek V3.2 | Text | 164K | 160K | $0.56 | $1.68 |
77
77
  | DeepSeek V4 Flash | Text | 1.0M | 384K | $0.14 | $0.28 |
78
+ | DeepSeek V4 Flash 0731 | Text | 1.0M | 384K | $0.14 | $0.28 |
78
79
  | DeepSeek V4 Pro | Text | 1.0M | 384K | $1.74 | $3.48 |
79
- | DeepSeek V4 Pro (router) | Text | 1.0M | 1.0M | $1.74 | $3.48 |
80
- | Gemma 4 26B A4B IT | Text + Image | 262K | 0 | Free | Free |
81
- | Gemma 4 31B IT | Text + Image | 262K | 0 | Free | Free |
80
+ | DeepSeek V4 Pro (router) | Text | 1.0M | 384K | $1.74 | $3.48 |
81
+ | Gemma 4 26B A4B IT | Text + Image | 262K | 33K | Free | Free |
82
+ | Gemma 4 31B IT | Text + Image | 262K | 33K | Free | Free |
82
83
  | GLM 4.5 | Text | 131K | 131K | $0.55 | $2.19 |
83
84
  | GLM 4.5 Air | Text | 131K | 131K | $0.22 | $0.88 |
84
85
  | GLM 4.7 | Text | 203K | 198K | $0.60 | $2.20 |
@@ -90,7 +91,7 @@ pi
90
91
  | GLM 5.2 Fast | Text | 1.0M | 131K | $2.10 | $6.60 |
91
92
  | GPT OSS 120B | Text | 131K | 33K | $0.15 | $0.60 |
92
93
  | GPT OSS 20B | Text | 131K | 33K | $0.07 | $0.30 |
93
- | Inkling | Text + Image | 1.0M | 0 | Free | Free |
94
+ | Inkling | Text + Image | 1.0M | 262K | $1.00 | $4.05 |
94
95
  | Kimi K2 Instruct | Text | 131K | 16K | $1.00 | $3.00 |
95
96
  | Kimi K2 Thinking | Text | 262K | 256K | $0.60 | $2.50 |
96
97
  | Kimi K2.5 | Text + Image | 262K | 256K | $0.60 | $3.00 |
@@ -101,18 +102,19 @@ pi
101
102
  | Kimi K2.6 Turbo | Text + Image | 262K | 262K | $2.00 | $8.00 |
102
103
  | Kimi K2.7 Code | Text + Image | 262K | 262K | $0.95 | $4.00 |
103
104
  | Kimi K2.7 Code Fast | Text + Image | 262K | 262K | $1.90 | $8.00 |
104
- | Kimi K3 | Text + Image | 1.0M | 0 | Free | Free |
105
- | Llama 3.3 70B Instruct | Text | 131K | 0 | Free | Free |
106
- | MiniMax M2.7 (router) | Text | 204K | 0 | $0.30 | $1.20 |
105
+ | Kimi K3 | Text + Image | 1.0M | 131K | $3.00 | $15.00 |
106
+ | Kimi K3 Fast | Text + Image | 1.0M | 131K | $4.50 | $22.50 |
107
+ | Llama 3.3 70B Instruct | Text | 131K | 4K | $0.90 | $0.90 |
108
+ | MiniMax M2.7 (router) | Text | 204K | 131K | $0.30 | $1.20 |
107
109
  | MiniMax-M2.1 | Text | 197K | 200K | $0.30 | $1.20 |
108
110
  | MiniMax-M2.5 | Text | 197K | 197K | $0.30 | $1.20 |
109
111
  | MiniMax-M2.7 | Text | 197K | 197K | $0.30 | $1.20 |
110
- | MiniMax-M3 | Text | 512K | 512K | $0.30 | $1.20 |
111
- | NVIDIA Nemotron 3 Ultra NVFP4 | Text | 262K | 0 | Free | Free |
112
+ | MiniMax-M3 | Text + Image | 512K | 512K | $0.30 | $1.20 |
113
+ | NVIDIA Nemotron 3 Ultra NVFP4 | Text | 262K | 66K | $0.60 | $2.40 |
112
114
  | Qwen 3.7 Plus | Text + Image | 262K | 66K | $0.40 | $1.60 |
113
- | Qwen3 8B | Text | 41K | 0 | Free | Free |
114
- | Qwen3 VL 30B A3B Instruct | Text + Image | 262K | 0 | Free | Free |
115
- | Qwen3 VL 30B A3B Thinking | Text + Image | 262K | 0 | Free | Free |
115
+ | Qwen3 8B | Text | 41K | 41K | $0.20 | $0.20 |
116
+ | Qwen3 VL 30B A3B Instruct | Text + Image | 262K | 33K | $0.50 | $0.50 |
117
+ | Qwen3 VL 30B A3B Thinking | Text + Image | 262K | 33K | $0.50 | $0.50 |
116
118
  *Costs are per million tokens. Prices subject to change - check [fireworks.ai](https://fireworks.ai) for current pricing.*
117
119
 
118
120
  ## Service Tiers
package/custom-models.json CHANGED
@@ -13,7 +13,7 @@
13
13
  "text"
14
14
  ],
15
15
  "contextWindow": 1048576,
16
- "maxTokens": 0
16
+ "maxTokens": 384000
17
17
  },
18
18
  {
19
19
  "id": "accounts/fireworks/routers/kimi-k2p5-fast",
@@ -30,7 +30,7 @@
30
30
  "image"
31
31
  ],
32
32
  "contextWindow": 262144,
33
- "maxTokens": 0
33
+ "maxTokens": 256000
34
34
  },
35
35
  {
36
36
  "id": "accounts/fireworks/routers/kimi-k2p6-turbo",
@@ -47,7 +47,7 @@
47
47
  "image"
48
48
  ],
49
49
  "contextWindow": 262144,
50
- "maxTokens": 0
50
+ "maxTokens": 262000
51
51
  },
52
52
  {
53
53
  "id": "accounts/fireworks/routers/kimi-k2p6",
@@ -96,7 +96,7 @@
96
96
  "text"
97
97
  ],
98
98
  "contextWindow": 204000,
99
- "maxTokens": 0
99
+ "maxTokens": 131072
100
100
  },
101
101
  {
102
102
  "id": "accounts/fireworks/routers/glm-5-fast",
@@ -129,7 +129,7 @@
129
129
  "image"
130
130
  ],
131
131
  "contextWindow": 262000,
132
- "maxTokens": 0
132
+ "maxTokens": 32768
133
133
  },
134
134
  {
135
135
  "id": "accounts/fireworks/models/gemma-4-26b-a4b-it",
@@ -146,7 +146,7 @@
146
146
  "image"
147
147
  ],
148
148
  "contextWindow": 262000,
149
- "maxTokens": 0
149
+ "maxTokens": 32768
150
150
  },
151
151
  {
152
152
  "id": "accounts/fireworks/models/qwen3p7-plus",
@@ -214,5 +214,34 @@
214
214
  ],
215
215
  "contextWindow": 262000,
216
216
  "maxTokens": 262000
217
+ },
218
+ {
219
+ "id": "accounts/fireworks/routers/kimi-k3-fast",
220
+ "name": "Kimi K3 Fast",
221
+ "reasoning": true,
222
+ "input": [
223
+ "text",
224
+ "image"
225
+ ],
226
+ "cost": {
227
+ "input": 4.5,
228
+ "output": 22.5,
229
+ "cacheRead": 0.45,
230
+ "cacheWrite": 0
231
+ },
232
+ "contextWindow": 1048576,
233
+ "maxTokens": 131072,
234
+ "thinkingLevelMap": {
235
+ "off": null,
236
+ "minimal": null,
237
+ "low": "low",
238
+ "medium": null,
239
+ "high": "high",
240
+ "xhigh": null,
241
+ "max": "max"
242
+ },
243
+ "compat": {
244
+ "supportsReasoningEffort": true
245
+ }
217
246
  }
218
247
  ]
package/models.json CHANGED
@@ -47,6 +47,22 @@
47
47
  "contextWindow": 1048576,
48
48
  "maxTokens": 0
49
49
  },
50
+ {
51
+ "id": "accounts/fireworks/models/deepseek-v4-flash-0731",
52
+ "name": "DeepSeek-V4-Flash-0731",
53
+ "reasoning": false,
54
+ "input": [
55
+ "text"
56
+ ],
57
+ "cost": {
58
+ "input": 0,
59
+ "output": 0,
60
+ "cacheRead": 0,
61
+ "cacheWrite": 0
62
+ },
63
+ "contextWindow": 1048576,
64
+ "maxTokens": 0
65
+ },
50
66
  {
51
67
  "id": "accounts/fireworks/models/deepseek-v4-pro",
52
68
  "name": "DeepSeek-V4-Pro",
package/package.json CHANGED
@@ -1,9 +1,18 @@
1
1
  {
2
2
  "name": "pi-fireworks-provider",
3
- "version": "1.5.2",
3
+ "version": "1.5.4",
4
4
  "description": "Fireworks AI provider extension for pi - Access Kimi, MiniMax, GLM, DeepSeek, and GPT-OSS models through the Fireworks AI API",
5
5
  "type": "module",
6
6
  "main": "index.ts",
7
+ "scripts": {
8
+ "clean": "echo 'nothing to clean'",
9
+ "build": "echo 'nothing to build'",
10
+ "check": "tsc --noEmit && vitest run && knip --no-gitignore",
11
+ "lint:dead": "knip --no-gitignore",
12
+ "test": "vitest run",
13
+ "test:watch": "vitest",
14
+ "update-models": "node scripts/update-models.js"
15
+ },
7
16
  "keywords": [
8
17
  "pi",
9
18
  "extension",
@@ -39,13 +48,5 @@
39
48
  "./index.ts"
40
49
  ]
41
50
  },
42
- "scripts": {
43
- "clean": "echo 'nothing to clean'",
44
- "build": "echo 'nothing to build'",
45
- "check": "tsc --noEmit && vitest run && knip --no-gitignore",
46
- "lint:dead": "knip --no-gitignore",
47
- "test": "vitest run",
48
- "test:watch": "vitest",
49
- "update-models": "node scripts/update-models.js"
50
- }
51
- }
51
+ "packageManager": "pnpm@11.18.0"
52
+ }
package/patch.json CHANGED
@@ -2,7 +2,9 @@
2
2
  "accounts/fireworks/models/deepseek-v3p1": {
3
3
  "name": "DeepSeek V3.1",
4
4
  "reasoning": true,
5
- "input": ["text"],
5
+ "input": [
6
+ "text"
7
+ ],
6
8
  "cost": {
7
9
  "input": 0.56,
8
10
  "output": 1.68,
@@ -18,7 +20,9 @@
18
20
  "accounts/fireworks/models/deepseek-v3p2": {
19
21
  "name": "DeepSeek V3.2",
20
22
  "reasoning": true,
21
- "input": ["text"],
23
+ "input": [
24
+ "text"
25
+ ],
22
26
  "cost": {
23
27
  "input": 0.56,
24
28
  "output": 1.68,
@@ -34,14 +38,9 @@
34
38
  "accounts/fireworks/models/deepseek-v4-flash": {
35
39
  "name": "DeepSeek V4 Flash",
36
40
  "reasoning": true,
37
- "thinkingLevelMap": {
38
- "off": "none",
39
- "minimal": null,
40
- "low": "high",
41
- "medium": "high",
42
- "max": "max"
43
- },
44
- "input": ["text"],
41
+ "input": [
42
+ "text"
43
+ ],
45
44
  "cost": {
46
45
  "input": 0.14,
47
46
  "output": 0.28,
@@ -57,14 +56,9 @@
57
56
  "accounts/fireworks/models/deepseek-v4-pro": {
58
57
  "name": "DeepSeek V4 Pro",
59
58
  "reasoning": true,
60
- "thinkingLevelMap": {
61
- "off": "none",
62
- "minimal": null,
63
- "low": "high",
64
- "medium": "high",
65
- "max": "max"
66
- },
67
- "input": ["text"],
59
+ "input": [
60
+ "text"
61
+ ],
68
62
  "cost": {
69
63
  "input": 1.74,
70
64
  "output": 3.48,
@@ -80,7 +74,9 @@
80
74
  "accounts/fireworks/models/glm-4p5": {
81
75
  "name": "GLM 4.5",
82
76
  "reasoning": true,
83
- "input": ["text"],
77
+ "input": [
78
+ "text"
79
+ ],
84
80
  "cost": {
85
81
  "input": 0.55,
86
82
  "output": 2.19,
@@ -96,7 +92,9 @@
96
92
  "accounts/fireworks/models/glm-4p5-air": {
97
93
  "name": "GLM 4.5 Air",
98
94
  "reasoning": true,
99
- "input": ["text"],
95
+ "input": [
96
+ "text"
97
+ ],
100
98
  "cost": {
101
99
  "input": 0.22,
102
100
  "output": 0.88,
@@ -112,7 +110,9 @@
112
110
  "accounts/fireworks/models/glm-4p7": {
113
111
  "name": "GLM 4.7",
114
112
  "reasoning": true,
115
- "input": ["text"],
113
+ "input": [
114
+ "text"
115
+ ],
116
116
  "cost": {
117
117
  "input": 0.6,
118
118
  "output": 2.2,
@@ -128,7 +128,9 @@
128
128
  "accounts/fireworks/models/glm-5": {
129
129
  "name": "GLM 5",
130
130
  "reasoning": true,
131
- "input": ["text"],
131
+ "input": [
132
+ "text"
133
+ ],
132
134
  "cost": {
133
135
  "input": 1,
134
136
  "output": 3.2,
@@ -144,7 +146,9 @@
144
146
  "accounts/fireworks/models/glm-5p1": {
145
147
  "name": "GLM 5.1",
146
148
  "reasoning": true,
147
- "input": ["text"],
149
+ "input": [
150
+ "text"
151
+ ],
148
152
  "cost": {
149
153
  "input": 1.4,
150
154
  "output": 4.4,
@@ -167,13 +171,17 @@
167
171
  "minimal": null,
168
172
  "low": "high",
169
173
  "medium": "high",
174
+ "high": "high",
175
+ "xhigh": null,
170
176
  "max": "max"
171
177
  },
172
- "input": ["text"],
178
+ "input": [
179
+ "text"
180
+ ],
173
181
  "cost": {
174
182
  "input": 1.4,
175
183
  "output": 4.4,
176
- "cacheRead": 0.26,
184
+ "cacheRead": 0.14,
177
185
  "cacheWrite": 0
178
186
  },
179
187
  "contextWindow": 1048575,
@@ -186,7 +194,9 @@
186
194
  "accounts/fireworks/models/gpt-oss-120b": {
187
195
  "name": "GPT OSS 120B",
188
196
  "reasoning": true,
189
- "input": ["text"],
197
+ "input": [
198
+ "text"
199
+ ],
190
200
  "cost": {
191
201
  "input": 0.15,
192
202
  "output": 0.6,
@@ -202,7 +212,9 @@
202
212
  "accounts/fireworks/models/gpt-oss-20b": {
203
213
  "name": "GPT OSS 20B",
204
214
  "reasoning": true,
205
- "input": ["text"],
215
+ "input": [
216
+ "text"
217
+ ],
206
218
  "cost": {
207
219
  "input": 0.07,
208
220
  "output": 0.3,
@@ -218,7 +230,10 @@
218
230
  "accounts/fireworks/models/gemma-4-26b-a4b-it": {
219
231
  "name": "Gemma 4 26B A4B IT",
220
232
  "reasoning": true,
221
- "input": ["text", "image"],
233
+ "input": [
234
+ "text",
235
+ "image"
236
+ ],
222
237
  "cost": {
223
238
  "input": 0,
224
239
  "output": 0,
@@ -228,12 +243,16 @@
228
243
  "contextWindow": 262000,
229
244
  "compat": {
230
245
  "supportsReasoningEffort": true
231
- }
246
+ },
247
+ "maxTokens": 32768
232
248
  },
233
249
  "accounts/fireworks/models/gemma-4-31b-it": {
234
250
  "name": "Gemma 4 31B IT",
235
251
  "reasoning": true,
236
- "input": ["text", "image"],
252
+ "input": [
253
+ "text",
254
+ "image"
255
+ ],
237
256
  "cost": {
238
257
  "input": 0,
239
258
  "output": 0,
@@ -243,12 +262,15 @@
243
262
  "contextWindow": 262000,
244
263
  "compat": {
245
264
  "supportsReasoningEffort": true
246
- }
265
+ },
266
+ "maxTokens": 32768
247
267
  },
248
268
  "accounts/fireworks/models/kimi-k2-instruct": {
249
269
  "name": "Kimi K2 Instruct",
250
270
  "reasoning": false,
251
- "input": ["text"],
271
+ "input": [
272
+ "text"
273
+ ],
252
274
  "cost": {
253
275
  "input": 1,
254
276
  "output": 3,
@@ -261,7 +283,9 @@
261
283
  "accounts/fireworks/models/kimi-k2-thinking": {
262
284
  "name": "Kimi K2 Thinking",
263
285
  "reasoning": true,
264
- "input": ["text"],
286
+ "input": [
287
+ "text"
288
+ ],
265
289
  "cost": {
266
290
  "input": 0.6,
267
291
  "output": 2.5,
@@ -277,7 +301,10 @@
277
301
  "accounts/fireworks/models/kimi-k2p5": {
278
302
  "name": "Kimi K2.5",
279
303
  "reasoning": true,
280
- "input": ["text", "image"],
304
+ "input": [
305
+ "text",
306
+ "image"
307
+ ],
281
308
  "cost": {
282
309
  "input": 0.6,
283
310
  "output": 3,
@@ -293,7 +320,10 @@
293
320
  "accounts/fireworks/models/kimi-k2p6": {
294
321
  "name": "Kimi K2.6",
295
322
  "reasoning": true,
296
- "input": ["text", "image"],
323
+ "input": [
324
+ "text",
325
+ "image"
326
+ ],
297
327
  "cost": {
298
328
  "input": 0.95,
299
329
  "output": 4,
@@ -309,7 +339,10 @@
309
339
  "accounts/fireworks/models/kimi-k2p7-code": {
310
340
  "name": "Kimi K2.7 Code",
311
341
  "reasoning": true,
312
- "input": ["text", "image"],
342
+ "input": [
343
+ "text",
344
+ "image"
345
+ ],
313
346
  "cost": {
314
347
  "input": 0.95,
315
348
  "output": 4,
@@ -325,7 +358,9 @@
325
358
  "accounts/fireworks/models/minimax-m2p1": {
326
359
  "name": "MiniMax-M2.1",
327
360
  "reasoning": true,
328
- "input": ["text"],
361
+ "input": [
362
+ "text"
363
+ ],
329
364
  "cost": {
330
365
  "input": 0.3,
331
366
  "output": 1.2,
@@ -341,7 +376,9 @@
341
376
  "accounts/fireworks/models/minimax-m2p5": {
342
377
  "name": "MiniMax-M2.5",
343
378
  "reasoning": true,
344
- "input": ["text"],
379
+ "input": [
380
+ "text"
381
+ ],
345
382
  "cost": {
346
383
  "input": 0.3,
347
384
  "output": 1.2,
@@ -357,7 +394,9 @@
357
394
  "accounts/fireworks/models/minimax-m2p7": {
358
395
  "name": "MiniMax-M2.7",
359
396
  "reasoning": true,
360
- "input": ["text"],
397
+ "input": [
398
+ "text"
399
+ ],
361
400
  "cost": {
362
401
  "input": 0.3,
363
402
  "output": 1.2,
@@ -373,7 +412,10 @@
373
412
  "accounts/fireworks/models/minimax-m3": {
374
413
  "name": "MiniMax-M3",
375
414
  "reasoning": true,
376
- "input": ["text"],
415
+ "input": [
416
+ "text",
417
+ "image"
418
+ ],
377
419
  "cost": {
378
420
  "input": 0.3,
379
421
  "output": 1.2,
@@ -389,7 +431,10 @@
389
431
  "accounts/fireworks/models/qwen3p7-plus": {
390
432
  "name": "Qwen 3.7 Plus",
391
433
  "reasoning": true,
392
- "input": ["text", "image"],
434
+ "input": [
435
+ "text",
436
+ "image"
437
+ ],
393
438
  "cost": {
394
439
  "input": 0.4,
395
440
  "output": 1.6,
@@ -412,7 +457,9 @@
412
457
  "medium": "high",
413
458
  "max": "max"
414
459
  },
415
- "input": ["text"],
460
+ "input": [
461
+ "text"
462
+ ],
416
463
  "cost": {
417
464
  "input": 1.74,
418
465
  "output": 3.48,
@@ -420,7 +467,7 @@
420
467
  "cacheWrite": 0
421
468
  },
422
469
  "contextWindow": 1048576,
423
- "maxTokens": 1048576,
470
+ "maxTokens": 384000,
424
471
  "compat": {
425
472
  "supportsReasoningEffort": true
426
473
  }
@@ -428,7 +475,9 @@
428
475
  "accounts/fireworks/routers/glm-5-fast": {
429
476
  "name": "GLM 5 Fast",
430
477
  "reasoning": true,
431
- "input": ["text"],
478
+ "input": [
479
+ "text"
480
+ ],
432
481
  "cost": {
433
482
  "input": 1,
434
483
  "output": 3.2,
@@ -444,7 +493,9 @@
444
493
  "accounts/fireworks/routers/glm-5p1-fast": {
445
494
  "name": "GLM 5.1 Fast",
446
495
  "reasoning": true,
447
- "input": ["text"],
496
+ "input": [
497
+ "text"
498
+ ],
448
499
  "cost": {
449
500
  "input": 2.8,
450
501
  "output": 8.8,
@@ -465,9 +516,13 @@
465
516
  "minimal": null,
466
517
  "low": "high",
467
518
  "medium": "high",
519
+ "high": "high",
520
+ "xhigh": null,
468
521
  "max": "max"
469
522
  },
470
- "input": ["text"],
523
+ "input": [
524
+ "text"
525
+ ],
471
526
  "cost": {
472
527
  "input": 2.1,
473
528
  "output": 6.6,
@@ -483,7 +538,10 @@
483
538
  "accounts/fireworks/routers/kimi-k2p5-fast": {
484
539
  "name": "Kimi K2.5 Fast (router)",
485
540
  "reasoning": true,
486
- "input": ["text", "image"],
541
+ "input": [
542
+ "text",
543
+ "image"
544
+ ],
487
545
  "cost": {
488
546
  "input": 0.6,
489
547
  "output": 3,
@@ -499,7 +557,10 @@
499
557
  "accounts/fireworks/routers/kimi-k2p6": {
500
558
  "name": "Kimi K2.6 (router)",
501
559
  "reasoning": true,
502
- "input": ["text", "image"],
560
+ "input": [
561
+ "text",
562
+ "image"
563
+ ],
503
564
  "cost": {
504
565
  "input": 0.95,
505
566
  "output": 4,
@@ -515,7 +576,10 @@
515
576
  "accounts/fireworks/routers/kimi-k2p6-fast": {
516
577
  "name": "Kimi K2.6 Fast",
517
578
  "reasoning": true,
518
- "input": ["text", "image"],
579
+ "input": [
580
+ "text",
581
+ "image"
582
+ ],
519
583
  "cost": {
520
584
  "input": 2,
521
585
  "output": 8,
@@ -531,7 +595,10 @@
531
595
  "accounts/fireworks/routers/kimi-k2p6-turbo": {
532
596
  "name": "Kimi K2.6 Turbo",
533
597
  "reasoning": true,
534
- "input": ["text", "image"],
598
+ "input": [
599
+ "text",
600
+ "image"
601
+ ],
535
602
  "cost": {
536
603
  "input": 2,
537
604
  "output": 8,
@@ -547,7 +614,10 @@
547
614
  "accounts/fireworks/routers/kimi-k2p7-code-fast": {
548
615
  "name": "Kimi K2.7 Code Fast",
549
616
  "reasoning": true,
550
- "input": ["text", "image"],
617
+ "input": [
618
+ "text",
619
+ "image"
620
+ ],
551
621
  "cost": {
552
622
  "input": 1.9,
553
623
  "output": 8,
@@ -563,7 +633,9 @@
563
633
  "accounts/fireworks/routers/minimax-m2p7": {
564
634
  "name": "MiniMax M2.7 (router)",
565
635
  "reasoning": true,
566
- "input": ["text"],
636
+ "input": [
637
+ "text"
638
+ ],
567
639
  "cost": {
568
640
  "input": 0.3,
569
641
  "output": 1.2,
@@ -571,6 +643,165 @@
571
643
  "cacheWrite": 0
572
644
  },
573
645
  "contextWindow": 204000,
646
+ "compat": {
647
+ "supportsReasoningEffort": true
648
+ },
649
+ "maxTokens": 131072
650
+ },
651
+ "accounts/fireworks/models/deepseek-v4-flash-0731": {
652
+ "name": "DeepSeek V4 Flash 0731",
653
+ "reasoning": true,
654
+ "thinkingLevelMap": {
655
+ "off": "none",
656
+ "minimal": null,
657
+ "low": "high",
658
+ "medium": "high",
659
+ "max": "max"
660
+ },
661
+ "input": [
662
+ "text"
663
+ ],
664
+ "cost": {
665
+ "input": 0.14,
666
+ "output": 0.28,
667
+ "cacheRead": 0.028,
668
+ "cacheWrite": 0
669
+ },
670
+ "contextWindow": 1000000,
671
+ "maxTokens": 384000,
672
+ "compat": {
673
+ "supportsReasoningEffort": true
674
+ }
675
+ },
676
+ "accounts/fireworks/models/inkling": {
677
+ "name": "Inkling",
678
+ "reasoning": true,
679
+ "input": [
680
+ "text",
681
+ "image"
682
+ ],
683
+ "cost": {
684
+ "input": 1,
685
+ "output": 4.05,
686
+ "cacheRead": 0.17,
687
+ "cacheWrite": 0
688
+ },
689
+ "contextWindow": 1048576,
690
+ "maxTokens": 262144,
691
+ "compat": {
692
+ "supportsReasoningEffort": true
693
+ }
694
+ },
695
+ "accounts/fireworks/models/kimi-k3": {
696
+ "name": "Kimi K3",
697
+ "reasoning": true,
698
+ "input": [
699
+ "text",
700
+ "image"
701
+ ],
702
+ "cost": {
703
+ "input": 3,
704
+ "output": 15,
705
+ "cacheRead": 0.3,
706
+ "cacheWrite": 0
707
+ },
708
+ "contextWindow": 1048576,
709
+ "maxTokens": 131072,
710
+ "thinkingLevelMap": {
711
+ "off": null,
712
+ "minimal": null,
713
+ "low": "low",
714
+ "medium": null,
715
+ "high": "high",
716
+ "xhigh": null,
717
+ "max": "max"
718
+ },
719
+ "compat": {
720
+ "supportsReasoningEffort": true
721
+ }
722
+ },
723
+ "accounts/fireworks/models/llama-v3p3-70b-instruct": {
724
+ "name": "Llama 3.3 70B Instruct",
725
+ "reasoning": false,
726
+ "input": [
727
+ "text"
728
+ ],
729
+ "cost": {
730
+ "input": 0.9,
731
+ "output": 0.9,
732
+ "cacheRead": 0,
733
+ "cacheWrite": 0
734
+ },
735
+ "contextWindow": 131072,
736
+ "maxTokens": 4096
737
+ },
738
+ "accounts/fireworks/models/nemotron-3-ultra-nvfp4": {
739
+ "name": "NVIDIA Nemotron 3 Ultra NVFP4",
740
+ "reasoning": true,
741
+ "input": [
742
+ "text"
743
+ ],
744
+ "cost": {
745
+ "input": 0.6,
746
+ "output": 2.4,
747
+ "cacheRead": 0.12,
748
+ "cacheWrite": 0
749
+ },
750
+ "contextWindow": 262144,
751
+ "maxTokens": 65536,
752
+ "compat": {
753
+ "supportsReasoningEffort": true
754
+ }
755
+ },
756
+ "accounts/fireworks/models/qwen3-8b": {
757
+ "name": "Qwen3 8B",
758
+ "reasoning": true,
759
+ "input": [
760
+ "text"
761
+ ],
762
+ "cost": {
763
+ "input": 0.2,
764
+ "output": 0.2,
765
+ "cacheRead": 0,
766
+ "cacheWrite": 0
767
+ },
768
+ "contextWindow": 40960,
769
+ "maxTokens": 40960,
770
+ "compat": {
771
+ "supportsReasoningEffort": true
772
+ }
773
+ },
774
+ "accounts/fireworks/models/qwen3-vl-30b-a3b-instruct": {
775
+ "name": "Qwen3 VL 30B A3B Instruct",
776
+ "reasoning": false,
777
+ "input": [
778
+ "text",
779
+ "image"
780
+ ],
781
+ "cost": {
782
+ "input": 0.5,
783
+ "output": 0.5,
784
+ "cacheRead": 0,
785
+ "cacheWrite": 0
786
+ },
787
+ "contextWindow": 262144,
788
+ "maxTokens": 32768
789
+ },
790
+ "accounts/fireworks/models/qwen3-vl-30b-a3b-thinking": {
791
+ "name": "Qwen3 VL 30B A3B Thinking",
792
+ "reasoning": true,
793
+ "input": [
794
+ "text",
795
+ "image"
796
+ ],
797
+ "cost": {
798
+ "input": 0.5,
799
+ "output": 0.5,
800
+ "cacheRead": 0,
801
+ "cacheWrite": 0
802
+ },
803
+ "contextWindow": 262144,
804
+ "maxTokens": 32768,
574
805
  "compat": {
575
806
  "supportsReasoningEffort": true
576
807
  }