pi-free 2.2.5 → 2.2.6

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 (30) hide show
  1. package/CHANGELOG.md +16 -0
  2. package/README.md +1 -1
  3. package/banner.svg +1 -1
  4. package/config.ts +803 -779
  5. package/constants.ts +0 -3
  6. package/index.ts +30 -24
  7. package/lib/built-in-toggle.ts +426 -427
  8. package/lib/provider-cache.ts +22 -0
  9. package/package.json +1 -1
  10. package/provider-failover/benchmark-lookup.ts +4 -1
  11. package/provider-failover/benchmarks-chunk-0.ts +570 -570
  12. package/provider-failover/benchmarks-chunk-1.ts +676 -676
  13. package/provider-failover/benchmarks-chunk-2.ts +673 -673
  14. package/provider-failover/benchmarks-chunk-3.ts +680 -680
  15. package/provider-failover/benchmarks-chunk-4.ts +683 -683
  16. package/provider-failover/benchmarks-chunk-5.ts +816 -474
  17. package/provider-helper.ts +64 -0
  18. package/providers/bai/bai.ts +8 -2
  19. package/providers/crofai/crofai.ts +8 -2
  20. package/providers/deepinfra/deepinfra.ts +8 -2
  21. package/providers/dynamic-built-in/index.ts +36 -26
  22. package/providers/kilo/kilo.ts +41 -22
  23. package/providers/novita/novita.ts +8 -2
  24. package/providers/openmodel/openmodel.ts +8 -2
  25. package/providers/routeway/routeway.ts +8 -2
  26. package/providers/sambanova/sambanova.ts +12 -6
  27. package/providers/together/together.ts +8 -2
  28. package/providers/tokenrouter/tokenrouter.ts +8 -2
  29. package/providers/zenmux/zenmux.ts +8 -2
  30. package/providers/codestral/codestral.ts +0 -128
package/provider-failover/benchmarks-chunk-0.ts CHANGED
@@ -1,20 +1,20 @@
1
1
  // Auto-generated benchmark data chunk 0
2
- // Models: gpt-oss-20b-low .. lfm2-8b-a1b (90 entries)
3
- // Last updated: 2026-06-01
2
+ // Models: gpt-oss-120b-high .. nova-2.0-lite-medium (90 entries)
3
+ // Last updated: 2026-07-01
4
4
  // DO NOT EDIT MANUALLY — generated by scripts/update-benchmarks.ts
5
5
 
6
6
  import type { HardcodedBenchmark } from "./hardcoded-benchmarks.ts";
7
7
 
8
8
  export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
9
- "gpt-oss-20b-low": {
9
+ "gpt-oss-120b-high": {
10
10
  // AA specific benchmarks
11
- codingIndex: 14.4,
12
- mathIndex: 62.3,
11
+ codingIndex: 30.4,
12
+ mathIndex: 93.4,
13
13
 
14
14
  // Academic benchmarks
15
- mmluPro: 0.718,
16
- gpqa: 0.611,
17
- hle: 0.051,
15
+ mmluPro: 0.808,
16
+ gpqa: 0.782,
17
+ hle: 0.185,
18
18
 
19
19
  // Capabilities
20
20
  contextWindow: 8192,
@@ -22,12 +22,12 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
22
22
  supportsVision: false,
23
23
 
24
24
  // Metadata
25
- lastUpdated: "2026-06-01",
26
- originalModel: "gpt-oss-20B (low)",
25
+ lastUpdated: "2026-07-01",
26
+ originalModel: "gpt-oss-120b (high)",
27
27
  },
28
28
  "gpt-oss-20b-high": {
29
29
  // AA specific benchmarks
30
- codingIndex: 18.5,
30
+ codingIndex: 20.7,
31
31
  mathIndex: 89.3,
32
32
 
33
33
  // Academic benchmarks
@@ -41,18 +41,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
41
41
  supportsVision: false,
42
42
 
43
43
  // Metadata
44
- lastUpdated: "2026-06-01",
44
+ lastUpdated: "2026-07-01",
45
45
  originalModel: "gpt-oss-20B (high)",
46
46
  },
47
- "gpt-oss-120b-high": {
47
+ "gpt-oss-20b-low": {
48
48
  // AA specific benchmarks
49
- codingIndex: 28.6,
50
- mathIndex: 93.4,
49
+ codingIndex: undefined,
50
+ mathIndex: 62.3,
51
51
 
52
52
  // Academic benchmarks
53
- mmluPro: 0.808,
54
- gpqa: 0.782,
55
- hle: 0.185,
53
+ mmluPro: 0.718,
54
+ gpqa: 0.611,
55
+ hle: 0.051,
56
56
 
57
57
  // Capabilities
58
58
  contextWindow: 8192,
@@ -60,18 +60,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
60
60
  supportsVision: false,
61
61
 
62
62
  // Metadata
63
- lastUpdated: "2026-06-01",
64
- originalModel: "gpt-oss-120b (high)",
63
+ lastUpdated: "2026-07-01",
64
+ originalModel: "gpt-oss-20B (low)",
65
65
  },
66
- "gpt-5.4-mini-xhigh": {
66
+ "gpt-5.5-low": {
67
67
  // AA specific benchmarks
68
- codingIndex: 51.5,
68
+ codingIndex: 60.9,
69
69
  mathIndex: undefined,
70
70
 
71
71
  // Academic benchmarks
72
72
  mmluPro: undefined,
73
- gpqa: 0.875,
74
- hle: 0.266,
73
+ gpqa: 0.91,
74
+ hle: 0.31,
75
75
 
76
76
  // Capabilities
77
77
  contextWindow: 8192,
@@ -79,18 +79,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
79
79
  supportsVision: false,
80
80
 
81
81
  // Metadata
82
- lastUpdated: "2026-06-01",
83
- originalModel: "GPT-5.4 mini (xhigh)",
82
+ lastUpdated: "2026-07-01",
83
+ originalModel: "GPT-5.5 (low)",
84
84
  },
85
- "gpt-5.4-nano-xhigh": {
85
+ "gpt-5.5-instant-june-2026": {
86
86
  // AA specific benchmarks
87
- codingIndex: 43.9,
87
+ codingIndex: 39.4,
88
88
  mathIndex: undefined,
89
89
 
90
90
  // Academic benchmarks
91
91
  mmluPro: undefined,
92
- gpqa: 0.817,
93
- hle: 0.265,
92
+ gpqa: 0.823,
93
+ hle: 0.186,
94
94
 
95
95
  // Capabilities
96
96
  contextWindow: 8192,
@@ -98,12 +98,12 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
98
98
  supportsVision: false,
99
99
 
100
100
  // Metadata
101
- lastUpdated: "2026-06-01",
102
- originalModel: "GPT-5.4 nano (xhigh)",
101
+ lastUpdated: "2026-07-01",
102
+ originalModel: "GPT-5.5 Instant (June 2026)",
103
103
  },
104
104
  "gpt-oss-120b-low": {
105
105
  // AA specific benchmarks
106
- codingIndex: 15.5,
106
+ codingIndex: undefined,
107
107
  mathIndex: 66.7,
108
108
 
109
109
  // Academic benchmarks
@@ -117,18 +117,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
117
117
  supportsVision: false,
118
118
 
119
119
  // Metadata
120
- lastUpdated: "2026-06-01",
120
+ lastUpdated: "2026-07-01",
121
121
  originalModel: "gpt-oss-120b (low)",
122
122
  },
123
- "o3": {
123
+ "gpt-5.4-nano-non-reasoning": {
124
124
  // AA specific benchmarks
125
- codingIndex: 38.4,
126
- mathIndex: 88.3,
125
+ codingIndex: undefined,
126
+ mathIndex: undefined,
127
127
 
128
128
  // Academic benchmarks
129
- mmluPro: 0.853,
130
- gpqa: 0.827,
131
- hle: 0.2,
129
+ mmluPro: undefined,
130
+ gpqa: 0.558,
131
+ hle: 0.042,
132
132
 
133
133
  // Capabilities
134
134
  contextWindow: 8192,
@@ -136,18 +136,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
136
136
  supportsVision: false,
137
137
 
138
138
  // Metadata
139
- lastUpdated: "2026-06-01",
140
- originalModel: "o3",
139
+ lastUpdated: "2026-07-01",
140
+ originalModel: "GPT-5.4 nano (Non-Reasoning)",
141
141
  },
142
- "grok-1": {
142
+ "gpt-5.4-mini-xhigh": {
143
143
  // AA specific benchmarks
144
- codingIndex: undefined,
144
+ codingIndex: 56.1,
145
145
  mathIndex: undefined,
146
146
 
147
147
  // Academic benchmarks
148
148
  mmluPro: undefined,
149
- gpqa: undefined,
150
- hle: undefined,
149
+ gpqa: 0.875,
150
+ hle: 0.266,
151
151
 
152
152
  // Capabilities
153
153
  contextWindow: 8192,
@@ -155,18 +155,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
155
155
  supportsVision: false,
156
156
 
157
157
  // Metadata
158
- lastUpdated: "2026-06-01",
159
- originalModel: "Grok-1",
158
+ lastUpdated: "2026-07-01",
159
+ originalModel: "GPT-5.4 mini (xhigh)",
160
160
  },
161
- "gpt-5.4-mini-medium": {
161
+ "grok-1": {
162
162
  // AA specific benchmarks
163
- codingIndex: 37.5,
163
+ codingIndex: undefined,
164
164
  mathIndex: undefined,
165
165
 
166
166
  // Academic benchmarks
167
167
  mmluPro: undefined,
168
- gpqa: 0.823,
169
- hle: 0.171,
168
+ gpqa: undefined,
169
+ hle: undefined,
170
170
 
171
171
  // Capabilities
172
172
  contextWindow: 8192,
@@ -174,18 +174,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
174
174
  supportsVision: false,
175
175
 
176
176
  // Metadata
177
- lastUpdated: "2026-06-01",
178
- originalModel: "GPT-5.4 mini (medium)",
177
+ lastUpdated: "2026-07-01",
178
+ originalModel: "Grok-1",
179
179
  },
180
- "gpt-5.4-mini-non-reasoning": {
180
+ "o3": {
181
181
  // AA specific benchmarks
182
- codingIndex: 25.3,
183
- mathIndex: undefined,
182
+ codingIndex: undefined,
183
+ mathIndex: 88.3,
184
184
 
185
185
  // Academic benchmarks
186
- mmluPro: undefined,
187
- gpqa: 0.606,
188
- hle: 0.057,
186
+ mmluPro: 0.853,
187
+ gpqa: 0.827,
188
+ hle: 0.2,
189
189
 
190
190
  // Capabilities
191
191
  contextWindow: 8192,
@@ -193,18 +193,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
193
193
  supportsVision: false,
194
194
 
195
195
  // Metadata
196
- lastUpdated: "2026-06-01",
197
- originalModel: "GPT-5.4 mini (Non-Reasoning)",
196
+ lastUpdated: "2026-07-01",
197
+ originalModel: "o3",
198
198
  },
199
- "gpt-5.5-instant-may-2026": {
199
+ "gpt-5.5-high": {
200
200
  // AA specific benchmarks
201
- codingIndex: 45.1,
201
+ codingIndex: 71.6,
202
202
  mathIndex: undefined,
203
203
 
204
204
  // Academic benchmarks
205
205
  mmluPro: undefined,
206
- gpqa: 0.846,
207
- hle: 0.203,
206
+ gpqa: 0.932,
207
+ hle: 0.43,
208
208
 
209
209
  // Capabilities
210
210
  contextWindow: 8192,
@@ -212,18 +212,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
212
212
  supportsVision: false,
213
213
 
214
214
  // Metadata
215
- lastUpdated: "2026-06-01",
216
- originalModel: "GPT-5.5 Instant (May 2026)",
215
+ lastUpdated: "2026-07-01",
216
+ originalModel: "GPT-5.5 (high)",
217
217
  },
218
- "gpt-5.5-low": {
218
+ "gpt-5.4-mini-non-reasoning": {
219
219
  // AA specific benchmarks
220
- codingIndex: 52.1,
220
+ codingIndex: undefined,
221
221
  mathIndex: undefined,
222
222
 
223
223
  // Academic benchmarks
224
224
  mmluPro: undefined,
225
- gpqa: 0.91,
226
- hle: 0.31,
225
+ gpqa: 0.606,
226
+ hle: 0.057,
227
227
 
228
228
  // Capabilities
229
229
  contextWindow: 8192,
@@ -231,18 +231,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
231
231
  supportsVision: false,
232
232
 
233
233
  // Metadata
234
- lastUpdated: "2026-06-01",
235
- originalModel: "GPT-5.5 (low)",
234
+ lastUpdated: "2026-07-01",
235
+ originalModel: "GPT-5.4 mini (Non-Reasoning)",
236
236
  },
237
- "gpt-5.4-nano-non-reasoning": {
237
+ "gpt-5.4-nano-medium": {
238
238
  // AA specific benchmarks
239
- codingIndex: 27.9,
239
+ codingIndex: undefined,
240
240
  mathIndex: undefined,
241
241
 
242
242
  // Academic benchmarks
243
243
  mmluPro: undefined,
244
- gpqa: 0.558,
245
- hle: 0.042,
244
+ gpqa: 0.761,
245
+ hle: 0.147,
246
246
 
247
247
  // Capabilities
248
248
  contextWindow: 8192,
@@ -250,18 +250,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
250
250
  supportsVision: false,
251
251
 
252
252
  // Metadata
253
- lastUpdated: "2026-06-01",
254
- originalModel: "GPT-5.4 nano (Non-Reasoning)",
253
+ lastUpdated: "2026-07-01",
254
+ originalModel: "GPT-5.4 nano (medium)",
255
255
  },
256
- "gpt-5.5-high": {
256
+ "gpt-5.5-non-reasoning": {
257
257
  // AA specific benchmarks
258
- codingIndex: 58.5,
258
+ codingIndex: 56.5,
259
259
  mathIndex: undefined,
260
260
 
261
261
  // Academic benchmarks
262
262
  mmluPro: undefined,
263
- gpqa: 0.932,
264
- hle: 0.43,
263
+ gpqa: 0.768,
264
+ hle: 0.126,
265
265
 
266
266
  // Capabilities
267
267
  contextWindow: 8192,
@@ -269,18 +269,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
269
269
  supportsVision: false,
270
270
 
271
271
  // Metadata
272
- lastUpdated: "2026-06-01",
273
- originalModel: "GPT-5.5 (high)",
272
+ lastUpdated: "2026-07-01",
273
+ originalModel: "GPT-5.5 (Non-reasoning)",
274
274
  },
275
- "gpt-5.4-nano-medium": {
275
+ "gpt-5.5-medium": {
276
276
  // AA specific benchmarks
277
- codingIndex: 35,
277
+ codingIndex: 71.5,
278
278
  mathIndex: undefined,
279
279
 
280
280
  // Academic benchmarks
281
281
  mmluPro: undefined,
282
- gpqa: 0.761,
283
- hle: 0.147,
282
+ gpqa: 0.926,
283
+ hle: 0.406,
284
284
 
285
285
  // Capabilities
286
286
  contextWindow: 8192,
@@ -288,18 +288,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
288
288
  supportsVision: false,
289
289
 
290
290
  // Metadata
291
- lastUpdated: "2026-06-01",
292
- originalModel: "GPT-5.4 nano (medium)",
291
+ lastUpdated: "2026-07-01",
292
+ originalModel: "GPT-5.5 (medium)",
293
293
  },
294
- "gpt-5.5-non-reasoning": {
294
+ "gpt-5.3-codex-xhigh": {
295
295
  // AA specific benchmarks
296
- codingIndex: 48.6,
296
+ codingIndex: undefined,
297
297
  mathIndex: undefined,
298
298
 
299
299
  // Academic benchmarks
300
300
  mmluPro: undefined,
301
- gpqa: 0.768,
302
- hle: 0.126,
301
+ gpqa: 0.915,
302
+ hle: 0.399,
303
303
 
304
304
  // Capabilities
305
305
  contextWindow: 8192,
@@ -307,18 +307,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
307
307
  supportsVision: false,
308
308
 
309
309
  // Metadata
310
- lastUpdated: "2026-06-01",
311
- originalModel: "GPT-5.5 (Non-reasoning)",
310
+ lastUpdated: "2026-07-01",
311
+ originalModel: "GPT-5.3 Codex (xhigh)",
312
312
  },
313
- "gpt-5.5-medium": {
313
+ "gpt-5.4-nano-xhigh": {
314
314
  // AA specific benchmarks
315
- codingIndex: 56.2,
315
+ codingIndex: 56.1,
316
316
  mathIndex: undefined,
317
317
 
318
318
  // Academic benchmarks
319
319
  mmluPro: undefined,
320
- gpqa: 0.926,
321
- hle: 0.406,
320
+ gpqa: 0.817,
321
+ hle: 0.265,
322
322
 
323
323
  // Capabilities
324
324
  contextWindow: 8192,
@@ -326,18 +326,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
326
326
  supportsVision: false,
327
327
 
328
328
  // Metadata
329
- lastUpdated: "2026-06-01",
330
- originalModel: "GPT-5.5 (medium)",
329
+ lastUpdated: "2026-07-01",
330
+ originalModel: "GPT-5.4 nano (xhigh)",
331
331
  },
332
- "gpt-5.3-codex-xhigh": {
332
+ "gpt-5.4-mini-medium": {
333
333
  // AA specific benchmarks
334
- codingIndex: 53.1,
334
+ codingIndex: undefined,
335
335
  mathIndex: undefined,
336
336
 
337
337
  // Academic benchmarks
338
338
  mmluPro: undefined,
339
- gpqa: 0.915,
340
- hle: 0.399,
339
+ gpqa: 0.823,
340
+ hle: 0.171,
341
341
 
342
342
  // Capabilities
343
343
  contextWindow: 8192,
@@ -345,12 +345,12 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
345
345
  supportsVision: false,
346
346
 
347
347
  // Metadata
348
- lastUpdated: "2026-06-01",
349
- originalModel: "GPT-5.3 Codex (xhigh)",
348
+ lastUpdated: "2026-07-01",
349
+ originalModel: "GPT-5.4 mini (medium)",
350
350
  },
351
351
  "gpt-5.5-xhigh": {
352
352
  // AA specific benchmarks
353
- codingIndex: 59.1,
353
+ codingIndex: 74.9,
354
354
  mathIndex: undefined,
355
355
 
356
356
  // Academic benchmarks
@@ -364,12 +364,12 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
364
364
  supportsVision: false,
365
365
 
366
366
  // Metadata
367
- lastUpdated: "2026-06-01",
367
+ lastUpdated: "2026-07-01",
368
368
  originalModel: "GPT-5.5 (xhigh)",
369
369
  },
370
370
  "llama-3.3-instruct-70b": {
371
371
  // AA specific benchmarks
372
- codingIndex: 10.7,
372
+ codingIndex: undefined,
373
373
  mathIndex: 7.7,
374
374
 
375
375
  // Academic benchmarks
@@ -383,12 +383,12 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
383
383
  supportsVision: false,
384
384
 
385
385
  // Metadata
386
- lastUpdated: "2026-06-01",
386
+ lastUpdated: "2026-07-01",
387
387
  originalModel: "Llama 3.3 Instruct 70B",
388
388
  },
389
389
  "llama-3.1-instruct-405b": {
390
390
  // AA specific benchmarks
391
- codingIndex: 14.5,
391
+ codingIndex: undefined,
392
392
  mathIndex: 3,
393
393
 
394
394
  // Academic benchmarks
@@ -402,7 +402,7 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
402
402
  supportsVision: false,
403
403
 
404
404
  // Metadata
405
- lastUpdated: "2026-06-01",
405
+ lastUpdated: "2026-07-01",
406
406
  originalModel: "Llama 3.1 Instruct 405B",
407
407
  },
408
408
  "llama-3.2-instruct-90b-vision": {
@@ -421,12 +421,12 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
421
421
  supportsVision: false,
422
422
 
423
423
  // Metadata
424
- lastUpdated: "2026-06-01",
424
+ lastUpdated: "2026-07-01",
425
425
  originalModel: "Llama 3.2 Instruct 90B (Vision)",
426
426
  },
427
427
  "llama-3.2-instruct-11b-vision": {
428
428
  // AA specific benchmarks
429
- codingIndex: 4.2,
429
+ codingIndex: undefined,
430
430
  mathIndex: 1.7,
431
431
 
432
432
  // Academic benchmarks
@@ -440,12 +440,31 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
440
440
  supportsVision: false,
441
441
 
442
442
  // Metadata
443
- lastUpdated: "2026-06-01",
443
+ lastUpdated: "2026-07-01",
444
444
  originalModel: "Llama 3.2 Instruct 11B (Vision)",
445
445
  },
446
+ "llama-4-scout": {
447
+ // AA specific benchmarks
448
+ codingIndex: 8.2,
449
+ mathIndex: 14,
450
+
451
+ // Academic benchmarks
452
+ mmluPro: 0.752,
453
+ gpqa: 0.587,
454
+ hle: 0.043,
455
+
456
+ // Capabilities
457
+ contextWindow: 8192,
458
+ supportsReasoning: false,
459
+ supportsVision: false,
460
+
461
+ // Metadata
462
+ lastUpdated: "2026-07-01",
463
+ originalModel: "Llama 4 Scout",
464
+ },
446
465
  "llama-4-maverick": {
447
466
  // AA specific benchmarks
448
- codingIndex: 15.6,
467
+ codingIndex: 16.3,
449
468
  mathIndex: 19.3,
450
469
 
451
470
  // Academic benchmarks
@@ -459,18 +478,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
459
478
  supportsVision: false,
460
479
 
461
480
  // Metadata
462
- lastUpdated: "2026-06-01",
481
+ lastUpdated: "2026-07-01",
463
482
  originalModel: "Llama 4 Maverick",
464
483
  },
465
- "llama-4-scout": {
484
+ "muse-spark": {
466
485
  // AA specific benchmarks
467
- codingIndex: 6.7,
468
- mathIndex: 14,
486
+ codingIndex: 58.6,
487
+ mathIndex: undefined,
469
488
 
470
489
  // Academic benchmarks
471
- mmluPro: 0.752,
472
- gpqa: 0.587,
473
- hle: 0.043,
490
+ mmluPro: undefined,
491
+ gpqa: 0.884,
492
+ hle: 0.399,
474
493
 
475
494
  // Capabilities
476
495
  contextWindow: 8192,
@@ -478,18 +497,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
478
497
  supportsVision: false,
479
498
 
480
499
  // Metadata
481
- lastUpdated: "2026-06-01",
482
- originalModel: "Llama 4 Scout",
500
+ lastUpdated: "2026-07-01",
501
+ originalModel: "Muse Spark",
483
502
  },
484
- "muse-spark": {
503
+ "gemma-4-e2b-reasoning": {
485
504
  // AA specific benchmarks
486
- codingIndex: 47.5,
505
+ codingIndex: undefined,
487
506
  mathIndex: undefined,
488
507
 
489
508
  // Academic benchmarks
490
509
  mmluPro: undefined,
491
- gpqa: 0.884,
492
- hle: 0.399,
510
+ gpqa: 0.433,
511
+ hle: 0.048,
493
512
 
494
513
  // Capabilities
495
514
  contextWindow: 8192,
@@ -497,18 +516,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
497
516
  supportsVision: false,
498
517
 
499
518
  // Metadata
500
- lastUpdated: "2026-06-01",
501
- originalModel: "Muse Spark",
519
+ lastUpdated: "2026-07-01",
520
+ originalModel: "Gemma 4 E2B (Reasoning)",
502
521
  },
503
- "gemini-3.1-pro-preview": {
522
+ "gemma-4-31b-non-reasoning": {
504
523
  // AA specific benchmarks
505
- codingIndex: 55.5,
524
+ codingIndex: undefined,
506
525
  mathIndex: undefined,
507
526
 
508
527
  // Academic benchmarks
509
528
  mmluPro: undefined,
510
- gpqa: 0.941,
511
- hle: 0.447,
529
+ gpqa: 0.763,
530
+ hle: 0.115,
512
531
 
513
532
  // Capabilities
514
533
  contextWindow: 8192,
@@ -516,18 +535,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
516
535
  supportsVision: false,
517
536
 
518
537
  // Metadata
519
- lastUpdated: "2026-06-01",
520
- originalModel: "Gemini 3.1 Pro Preview",
538
+ lastUpdated: "2026-07-01",
539
+ originalModel: "Gemma 4 31B (Non-reasoning)",
521
540
  },
522
- "gemma-4-26b-a4b-non-reasoning": {
541
+ "gemma-4-31b-reasoning": {
523
542
  // AA specific benchmarks
524
- codingIndex: 29.1,
543
+ codingIndex: 43.4,
525
544
  mathIndex: undefined,
526
545
 
527
546
  // Academic benchmarks
528
547
  mmluPro: undefined,
529
- gpqa: 0.714,
530
- hle: 0.107,
548
+ gpqa: 0.857,
549
+ hle: 0.227,
531
550
 
532
551
  // Capabilities
533
552
  contextWindow: 8192,
@@ -535,18 +554,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
535
554
  supportsVision: false,
536
555
 
537
556
  // Metadata
538
- lastUpdated: "2026-06-01",
539
- originalModel: "Gemma 4 26B A4B (Non-reasoning)",
557
+ lastUpdated: "2026-07-01",
558
+ originalModel: "Gemma 4 31B (Reasoning)",
540
559
  },
541
- "gemini-3.5-flash-medium": {
560
+ "gemma-4-12b-reasoning": {
542
561
  // AA specific benchmarks
543
- codingIndex: 43.9,
562
+ codingIndex: undefined,
544
563
  mathIndex: undefined,
545
564
 
546
565
  // Academic benchmarks
547
566
  mmluPro: undefined,
548
- gpqa: 0.921,
549
- hle: 0.399,
567
+ gpqa: 0.753,
568
+ hle: 0.148,
550
569
 
551
570
  // Capabilities
552
571
  contextWindow: 8192,
@@ -554,18 +573,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
554
573
  supportsVision: false,
555
574
 
556
575
  // Metadata
557
- lastUpdated: "2026-06-01",
558
- originalModel: "Gemini 3.5 Flash (medium)",
576
+ lastUpdated: "2026-07-01",
577
+ originalModel: "Gemma 4 12B (Reasoning)",
559
578
  },
560
- "gemma-4-e2b-non-reasoning": {
579
+ "gemma-4-26b-a4b-non-reasoning": {
561
580
  // AA specific benchmarks
562
- codingIndex: 8.3,
581
+ codingIndex: undefined,
563
582
  mathIndex: undefined,
564
583
 
565
584
  // Academic benchmarks
566
585
  mmluPro: undefined,
567
- gpqa: 0.405,
568
- hle: 0.045,
586
+ gpqa: 0.714,
587
+ hle: 0.107,
569
588
 
570
589
  // Capabilities
571
590
  contextWindow: 8192,
@@ -573,18 +592,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
573
592
  supportsVision: false,
574
593
 
575
594
  // Metadata
576
- lastUpdated: "2026-06-01",
577
- originalModel: "Gemma 4 E2B (Non-reasoning)",
595
+ lastUpdated: "2026-07-01",
596
+ originalModel: "Gemma 4 26B A4B (Non-reasoning)",
578
597
  },
579
- "gemma-4-31b-reasoning": {
598
+ "gemma-4-e4b-reasoning": {
580
599
  // AA specific benchmarks
581
- codingIndex: 38.7,
600
+ codingIndex: undefined,
582
601
  mathIndex: undefined,
583
602
 
584
603
  // Academic benchmarks
585
604
  mmluPro: undefined,
586
- gpqa: 0.857,
587
- hle: 0.227,
605
+ gpqa: 0.576,
606
+ hle: 0.037,
588
607
 
589
608
  // Capabilities
590
609
  contextWindow: 8192,
@@ -592,18 +611,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
592
611
  supportsVision: false,
593
612
 
594
613
  // Metadata
595
- lastUpdated: "2026-06-01",
596
- originalModel: "Gemma 4 31B (Reasoning)",
614
+ lastUpdated: "2026-07-01",
615
+ originalModel: "Gemma 4 E4B (Reasoning)",
597
616
  },
598
- "gemma-4-31b-non-reasoning": {
617
+ "diffusiongemma-26b-a4b": {
599
618
  // AA specific benchmarks
600
- codingIndex: 33.9,
619
+ codingIndex: 19.7,
601
620
  mathIndex: undefined,
602
621
 
603
622
  // Academic benchmarks
604
623
  mmluPro: undefined,
605
- gpqa: 0.763,
606
- hle: 0.115,
624
+ gpqa: 0.669,
625
+ hle: 0.102,
607
626
 
608
627
  // Capabilities
609
628
  contextWindow: 8192,
@@ -611,12 +630,12 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
611
630
  supportsVision: false,
612
631
 
613
632
  // Metadata
614
- lastUpdated: "2026-06-01",
615
- originalModel: "Gemma 4 31B (Non-reasoning)",
633
+ lastUpdated: "2026-07-01",
634
+ originalModel: "DiffusionGemma 26B A4B",
616
635
  },
617
636
  "gemma-4-26b-a4b-reasoning": {
618
637
  // AA specific benchmarks
619
- codingIndex: 22.4,
638
+ codingIndex: 39.3,
620
639
  mathIndex: undefined,
621
640
 
622
641
  // Academic benchmarks
@@ -630,18 +649,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
630
649
  supportsVision: false,
631
650
 
632
651
  // Metadata
633
- lastUpdated: "2026-06-01",
652
+ lastUpdated: "2026-07-01",
634
653
  originalModel: "Gemma 4 26B A4B (Reasoning)",
635
654
  },
636
- "gemini-3.1-flash-lite": {
655
+ "gemini-3.1-pro-preview": {
637
656
  // AA specific benchmarks
638
- codingIndex: 30.1,
657
+ codingIndex: 68.8,
639
658
  mathIndex: undefined,
640
659
 
641
660
  // Academic benchmarks
642
661
  mmluPro: undefined,
643
- gpqa: 0.822,
644
- hle: 0.162,
662
+ gpqa: 0.941,
663
+ hle: 0.447,
645
664
 
646
665
  // Capabilities
647
666
  contextWindow: 8192,
@@ -649,18 +668,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
649
668
  supportsVision: false,
650
669
 
651
670
  // Metadata
652
- lastUpdated: "2026-06-01",
653
- originalModel: "Gemini 3.1 Flash-Lite",
671
+ lastUpdated: "2026-07-01",
672
+ originalModel: "Gemini 3.1 Pro Preview",
654
673
  },
655
- "gemma-4-e4b-non-reasoning": {
674
+ "gemma-4-e2b-non-reasoning": {
656
675
  // AA specific benchmarks
657
- codingIndex: 6.4,
676
+ codingIndex: undefined,
658
677
  mathIndex: undefined,
659
678
 
660
679
  // Academic benchmarks
661
680
  mmluPro: undefined,
662
- gpqa: 0.549,
663
- hle: 0.047,
681
+ gpqa: 0.405,
682
+ hle: 0.045,
664
683
 
665
684
  // Capabilities
666
685
  contextWindow: 8192,
@@ -668,18 +687,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
668
687
  supportsVision: false,
669
688
 
670
689
  // Metadata
671
- lastUpdated: "2026-06-01",
672
- originalModel: "Gemma 4 E4B (Non-reasoning)",
690
+ lastUpdated: "2026-07-01",
691
+ originalModel: "Gemma 4 E2B (Non-reasoning)",
673
692
  },
674
- "gemma-3-270m": {
693
+ "gemma-4-12b-non-reasoning": {
675
694
  // AA specific benchmarks
676
- codingIndex: 0,
677
- mathIndex: 2.3,
695
+ codingIndex: undefined,
696
+ mathIndex: undefined,
678
697
 
679
698
  // Academic benchmarks
680
- mmluPro: 0.055,
681
- gpqa: 0.224,
682
- hle: 0.042,
699
+ mmluPro: undefined,
700
+ gpqa: 0.661,
701
+ hle: 0.062,
683
702
 
684
703
  // Capabilities
685
704
  contextWindow: 8192,
@@ -687,18 +706,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
687
706
  supportsVision: false,
688
707
 
689
708
  // Metadata
690
- lastUpdated: "2026-06-01",
691
- originalModel: "Gemma 3 270M",
709
+ lastUpdated: "2026-07-01",
710
+ originalModel: "Gemma 4 12B (Non-reasoning)",
692
711
  },
693
- "gemma-4-e4b-reasoning": {
712
+ "gemini-3.1-flash-lite": {
694
713
  // AA specific benchmarks
695
- codingIndex: 13.7,
714
+ codingIndex: 34.7,
696
715
  mathIndex: undefined,
697
716
 
698
717
  // Academic benchmarks
699
718
  mmluPro: undefined,
700
- gpqa: 0.576,
701
- hle: 0.037,
719
+ gpqa: 0.822,
720
+ hle: 0.162,
702
721
 
703
722
  // Capabilities
704
723
  contextWindow: 8192,
@@ -706,18 +725,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
706
725
  supportsVision: false,
707
726
 
708
727
  // Metadata
709
- lastUpdated: "2026-06-01",
710
- originalModel: "Gemma 4 E4B (Reasoning)",
728
+ lastUpdated: "2026-07-01",
729
+ originalModel: "Gemini 3.1 Flash-Lite",
711
730
  },
712
- "gemini-3.5-flash-high": {
731
+ "gemini-3.5-flash-medium": {
713
732
  // AA specific benchmarks
714
- codingIndex: 45,
733
+ codingIndex: undefined,
715
734
  mathIndex: undefined,
716
735
 
717
736
  // Academic benchmarks
718
737
  mmluPro: undefined,
719
- gpqa: 0.922,
720
- hle: 0.41,
738
+ gpqa: 0.921,
739
+ hle: 0.399,
721
740
 
722
741
  // Capabilities
723
742
  contextWindow: 8192,
@@ -725,18 +744,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
725
744
  supportsVision: false,
726
745
 
727
746
  // Metadata
728
- lastUpdated: "2026-06-01",
729
- originalModel: "Gemini 3.5 Flash (high)",
747
+ lastUpdated: "2026-07-01",
748
+ originalModel: "Gemini 3.5 Flash (medium)",
730
749
  },
731
- "gemini-3.5-flash-minimal": {
750
+ "gemma-4-e4b-non-reasoning": {
732
751
  // AA specific benchmarks
733
- codingIndex: 47.1,
752
+ codingIndex: undefined,
734
753
  mathIndex: undefined,
735
754
 
736
755
  // Academic benchmarks
737
756
  mmluPro: undefined,
738
- gpqa: 0.828,
739
- hle: 0.231,
757
+ gpqa: 0.549,
758
+ hle: 0.047,
740
759
 
741
760
  // Capabilities
742
761
  contextWindow: 8192,
@@ -744,18 +763,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
744
763
  supportsVision: false,
745
764
 
746
765
  // Metadata
747
- lastUpdated: "2026-06-01",
748
- originalModel: "Gemini 3.5 Flash (minimal)",
766
+ lastUpdated: "2026-07-01",
767
+ originalModel: "Gemma 4 E4B (Non-reasoning)",
749
768
  },
750
- "gemma-4-e2b-reasoning": {
769
+ "gemini-2.5-pro": {
751
770
  // AA specific benchmarks
752
- codingIndex: 9,
753
- mathIndex: undefined,
771
+ codingIndex: 33.3,
772
+ mathIndex: 87.7,
754
773
 
755
774
  // Academic benchmarks
756
- mmluPro: undefined,
757
- gpqa: 0.433,
758
- hle: 0.048,
775
+ mmluPro: 0.862,
776
+ gpqa: 0.844,
777
+ hle: 0.211,
759
778
 
760
779
  // Capabilities
761
780
  contextWindow: 8192,
@@ -763,18 +782,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
763
782
  supportsVision: false,
764
783
 
765
784
  // Metadata
766
- lastUpdated: "2026-06-01",
767
- originalModel: "Gemma 4 E2B (Reasoning)",
785
+ lastUpdated: "2026-07-01",
786
+ originalModel: "Gemini 2.5 Pro",
768
787
  },
769
- "gemini-2.5-pro": {
788
+ "gemini-3.5-flash-minimal": {
770
789
  // AA specific benchmarks
771
- codingIndex: 32,
772
- mathIndex: 87.7,
790
+ codingIndex: undefined,
791
+ mathIndex: undefined,
773
792
 
774
793
  // Academic benchmarks
775
- mmluPro: 0.862,
776
- gpqa: 0.844,
777
- hle: 0.211,
794
+ mmluPro: undefined,
795
+ gpqa: 0.828,
796
+ hle: 0.231,
778
797
 
779
798
  // Capabilities
780
799
  contextWindow: 8192,
@@ -782,18 +801,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
782
801
  supportsVision: false,
783
802
 
784
803
  // Metadata
785
- lastUpdated: "2026-06-01",
786
- originalModel: "Gemini 2.5 Pro",
804
+ lastUpdated: "2026-07-01",
805
+ originalModel: "Gemini 3.5 Flash (minimal)",
787
806
  },
788
- "claude-sonnet-4.6-non-reasoning-high-effort": {
807
+ "gemini-3.5-flash-high": {
789
808
  // AA specific benchmarks
790
- codingIndex: 46.4,
809
+ codingIndex: 70.1,
791
810
  mathIndex: undefined,
792
811
 
793
812
  // Academic benchmarks
794
813
  mmluPro: undefined,
795
- gpqa: 0.799,
796
- hle: 0.132,
814
+ gpqa: 0.922,
815
+ hle: 0.41,
797
816
 
798
817
  // Capabilities
799
818
  contextWindow: 8192,
@@ -801,18 +820,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
801
820
  supportsVision: false,
802
821
 
803
822
  // Metadata
804
- lastUpdated: "2026-06-01",
805
- originalModel: "Claude Sonnet 4.6 (Non-reasoning, High Effort)",
823
+ lastUpdated: "2026-07-01",
824
+ originalModel: "Gemini 3.5 Flash (high)",
806
825
  },
807
- "claude-opus-4.8-adaptive-reasoning-max-effort": {
826
+ "gemma-3-270m": {
808
827
  // AA specific benchmarks
809
- codingIndex: 56.7,
810
- mathIndex: undefined,
828
+ codingIndex: undefined,
829
+ mathIndex: 2.3,
811
830
 
812
831
  // Academic benchmarks
813
- mmluPro: undefined,
814
- gpqa: 0.92,
815
- hle: 0.457,
832
+ mmluPro: 0.055,
833
+ gpqa: 0.224,
834
+ hle: 0.042,
816
835
 
817
836
  // Capabilities
818
837
  contextWindow: 8192,
@@ -820,12 +839,12 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
820
839
  supportsVision: false,
821
840
 
822
841
  // Metadata
823
- lastUpdated: "2026-06-01",
824
- originalModel: "Claude Opus 4.8 (Adaptive Reasoning, Max Effort)",
842
+ lastUpdated: "2026-07-01",
843
+ originalModel: "Gemma 3 270M",
825
844
  },
826
845
  "claude-4.5-haiku-reasoning": {
827
846
  // AA specific benchmarks
828
- codingIndex: 32.6,
847
+ codingIndex: 43.9,
829
848
  mathIndex: 83.7,
830
849
 
831
850
  // Academic benchmarks
@@ -839,12 +858,31 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
839
858
  supportsVision: false,
840
859
 
841
860
  // Metadata
842
- lastUpdated: "2026-06-01",
861
+ lastUpdated: "2026-07-01",
843
862
  originalModel: "Claude 4.5 Haiku (Reasoning)",
844
863
  },
864
+ "claude-fable-5-adaptive-reasoning-max-effort-opus-4.8-fallback": {
865
+ // AA specific benchmarks
866
+ codingIndex: 76.5,
867
+ mathIndex: undefined,
868
+
869
+ // Academic benchmarks
870
+ mmluPro: undefined,
871
+ gpqa: 0.926,
872
+ hle: 0.533,
873
+
874
+ // Capabilities
875
+ contextWindow: 8192,
876
+ supportsReasoning: false,
877
+ supportsVision: false,
878
+
879
+ // Metadata
880
+ lastUpdated: "2026-07-01",
881
+ originalModel: "Claude Fable 5 (Adaptive Reasoning, Max Effort, Opus 4.8 Fallback)",
882
+ },
845
883
  "claude-opus-4.7-adaptive-reasoning-max-effort": {
846
884
  // AA specific benchmarks
847
- codingIndex: 52.5,
885
+ codingIndex: 73.6,
848
886
  mathIndex: undefined,
849
887
 
850
888
  // Academic benchmarks
@@ -858,18 +896,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
858
896
  supportsVision: false,
859
897
 
860
898
  // Metadata
861
- lastUpdated: "2026-06-01",
899
+ lastUpdated: "2026-07-01",
862
900
  originalModel: "Claude Opus 4.7 (Adaptive Reasoning, Max Effort)",
863
901
  },
864
- "claude-opus-4.7-non-reasoning-high-effort": {
902
+ "claude-sonnet-4.6-non-reasoning-high-effort": {
865
903
  // AA specific benchmarks
866
- codingIndex: 53.1,
904
+ codingIndex: undefined,
867
905
  mathIndex: undefined,
868
906
 
869
907
  // Academic benchmarks
870
908
  mmluPro: undefined,
871
- gpqa: 0.885,
872
- hle: 0.312,
909
+ gpqa: 0.799,
910
+ hle: 0.132,
873
911
 
874
912
  // Capabilities
875
913
  contextWindow: 8192,
@@ -877,18 +915,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
877
915
  supportsVision: false,
878
916
 
879
917
  // Metadata
880
- lastUpdated: "2026-06-01",
881
- originalModel: "Claude Opus 4.7 (Non-reasoning, High Effort)",
918
+ lastUpdated: "2026-07-01",
919
+ originalModel: "Claude Sonnet 4.6 (Non-reasoning, High Effort)",
882
920
  },
883
- "claude-sonnet-4.6-non-reasoning-low-effort": {
921
+ "claude-opus-4.8-adaptive-reasoning-max-effort": {
884
922
  // AA specific benchmarks
885
- codingIndex: 43,
923
+ codingIndex: 74.3,
886
924
  mathIndex: undefined,
887
925
 
888
926
  // Academic benchmarks
889
927
  mmluPro: undefined,
890
- gpqa: 0.797,
891
- hle: 0.108,
928
+ gpqa: 0.92,
929
+ hle: 0.457,
892
930
 
893
931
  // Capabilities
894
932
  contextWindow: 8192,
@@ -896,18 +934,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
896
934
  supportsVision: false,
897
935
 
898
936
  // Metadata
899
- lastUpdated: "2026-06-01",
900
- originalModel: "Claude Sonnet 4.6 (Non-reasoning, Low Effort)",
937
+ lastUpdated: "2026-07-01",
938
+ originalModel: "Claude Opus 4.8 (Adaptive Reasoning, Max Effort)",
901
939
  },
902
- "claude-4.5-haiku-non-reasoning": {
940
+ "claude-sonnet-5-adaptive-reasoning-max-effort": {
903
941
  // AA specific benchmarks
904
- codingIndex: 29.6,
905
- mathIndex: 39,
942
+ codingIndex: 71.5,
943
+ mathIndex: undefined,
906
944
 
907
945
  // Academic benchmarks
908
- mmluPro: 0.8,
909
- gpqa: 0.646,
910
- hle: 0.043,
946
+ mmluPro: undefined,
947
+ gpqa: 0.911,
948
+ hle: 0.396,
911
949
 
912
950
  // Capabilities
913
951
  contextWindow: 8192,
@@ -915,12 +953,12 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
915
953
  supportsVision: false,
916
954
 
917
955
  // Metadata
918
- lastUpdated: "2026-06-01",
919
- originalModel: "Claude 4.5 Haiku (Non-reasoning)",
956
+ lastUpdated: "2026-07-01",
957
+ originalModel: "Claude Sonnet 5 (Adaptive Reasoning, Max Effort)",
920
958
  },
921
959
  "claude-sonnet-4.6-adaptive-reasoning-max-effort": {
922
960
  // AA specific benchmarks
923
- codingIndex: 50.9,
961
+ codingIndex: 63,
924
962
  mathIndex: undefined,
925
963
 
926
964
  // Academic benchmarks
@@ -934,18 +972,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
934
972
  supportsVision: false,
935
973
 
936
974
  // Metadata
937
- lastUpdated: "2026-06-01",
975
+ lastUpdated: "2026-07-01",
938
976
  originalModel: "Claude Sonnet 4.6 (Adaptive Reasoning, Max Effort)",
939
977
  },
940
- "magistral-small-1.2": {
978
+ "claude-4.5-haiku-non-reasoning": {
941
979
  // AA specific benchmarks
942
- codingIndex: 14.8,
943
- mathIndex: 80.3,
980
+ codingIndex: undefined,
981
+ mathIndex: 39,
944
982
 
945
983
  // Academic benchmarks
946
- mmluPro: 0.768,
947
- gpqa: 0.663,
948
- hle: 0.061,
984
+ mmluPro: 0.8,
985
+ gpqa: 0.646,
986
+ hle: 0.043,
949
987
 
950
988
  // Capabilities
951
989
  contextWindow: 8192,
@@ -953,18 +991,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
953
991
  supportsVision: false,
954
992
 
955
993
  // Metadata
956
- lastUpdated: "2026-06-01",
957
- originalModel: "Magistral Small 1.2",
994
+ lastUpdated: "2026-07-01",
995
+ originalModel: "Claude 4.5 Haiku (Non-reasoning)",
958
996
  },
959
- "mistral-medium-3.5": {
997
+ "claude-sonnet-4.6-non-reasoning-low-effort": {
960
998
  // AA specific benchmarks
961
- codingIndex: 35.4,
999
+ codingIndex: undefined,
962
1000
  mathIndex: undefined,
963
1001
 
964
1002
  // Academic benchmarks
965
1003
  mmluPro: undefined,
966
- gpqa: 0.748,
967
- hle: 0.128,
1004
+ gpqa: 0.797,
1005
+ hle: 0.108,
968
1006
 
969
1007
  // Capabilities
970
1008
  contextWindow: 8192,
@@ -972,18 +1010,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
972
1010
  supportsVision: false,
973
1011
 
974
1012
  // Metadata
975
- lastUpdated: "2026-06-01",
976
- originalModel: "Mistral Medium 3.5",
1013
+ lastUpdated: "2026-07-01",
1014
+ originalModel: "Claude Sonnet 4.6 (Non-reasoning, Low Effort)",
977
1015
  },
978
- "devstral-small-2": {
1016
+ "claude-opus-4.7-non-reasoning-high-effort": {
979
1017
  // AA specific benchmarks
980
- codingIndex: 20.7,
981
- mathIndex: 34.3,
1018
+ codingIndex: undefined,
1019
+ mathIndex: undefined,
982
1020
 
983
1021
  // Academic benchmarks
984
- mmluPro: 0.678,
985
- gpqa: 0.532,
986
- hle: 0.034,
1022
+ mmluPro: undefined,
1023
+ gpqa: 0.885,
1024
+ hle: 0.312,
987
1025
 
988
1026
  // Capabilities
989
1027
  contextWindow: 8192,
@@ -991,18 +1029,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
991
1029
  supportsVision: false,
992
1030
 
993
1031
  // Metadata
994
- lastUpdated: "2026-06-01",
995
- originalModel: "Devstral Small 2",
1032
+ lastUpdated: "2026-07-01",
1033
+ originalModel: "Claude Opus 4.7 (Non-reasoning, High Effort)",
996
1034
  },
997
- "ministral-3-8b": {
1035
+ "mistral-large-3": {
998
1036
  // AA specific benchmarks
999
- codingIndex: 10,
1000
- mathIndex: 31.7,
1037
+ codingIndex: 20.1,
1038
+ mathIndex: 38,
1001
1039
 
1002
1040
  // Academic benchmarks
1003
- mmluPro: 0.642,
1004
- gpqa: 0.471,
1005
- hle: 0.043,
1041
+ mmluPro: 0.807,
1042
+ gpqa: 0.68,
1043
+ hle: 0.041,
1006
1044
 
1007
1045
  // Capabilities
1008
1046
  contextWindow: 8192,
@@ -1010,18 +1048,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
1010
1048
  supportsVision: false,
1011
1049
 
1012
1050
  // Metadata
1013
- lastUpdated: "2026-06-01",
1014
- originalModel: "Ministral 3 8B",
1051
+ lastUpdated: "2026-07-01",
1052
+ originalModel: "Mistral Large 3",
1015
1053
  },
1016
- "magistral-medium-1.2": {
1054
+ "mistral-small-4-reasoning": {
1017
1055
  // AA specific benchmarks
1018
- codingIndex: 21.7,
1019
- mathIndex: 82,
1056
+ codingIndex: undefined,
1057
+ mathIndex: undefined,
1020
1058
 
1021
1059
  // Academic benchmarks
1022
- mmluPro: 0.815,
1023
- gpqa: 0.739,
1024
- hle: 0.096,
1060
+ mmluPro: undefined,
1061
+ gpqa: 0.769,
1062
+ hle: 0.095,
1025
1063
 
1026
1064
  // Capabilities
1027
1065
  contextWindow: 8192,
@@ -1029,18 +1067,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
1029
1067
  supportsVision: false,
1030
1068
 
1031
1069
  // Metadata
1032
- lastUpdated: "2026-06-01",
1033
- originalModel: "Magistral Medium 1.2",
1070
+ lastUpdated: "2026-07-01",
1071
+ originalModel: "Mistral Small 4 (Reasoning)",
1034
1072
  },
1035
- "mistral-large-3": {
1073
+ "mistral-medium-3.5": {
1036
1074
  // AA specific benchmarks
1037
- codingIndex: 22.7,
1038
- mathIndex: 38,
1075
+ codingIndex: 46.9,
1076
+ mathIndex: undefined,
1039
1077
 
1040
1078
  // Academic benchmarks
1041
- mmluPro: 0.807,
1042
- gpqa: 0.68,
1043
- hle: 0.041,
1079
+ mmluPro: undefined,
1080
+ gpqa: 0.748,
1081
+ hle: 0.128,
1044
1082
 
1045
1083
  // Capabilities
1046
1084
  contextWindow: 8192,
@@ -1048,12 +1086,12 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
1048
1086
  supportsVision: false,
1049
1087
 
1050
1088
  // Metadata
1051
- lastUpdated: "2026-06-01",
1052
- originalModel: "Mistral Large 3",
1089
+ lastUpdated: "2026-07-01",
1090
+ originalModel: "Mistral Medium 3.5",
1053
1091
  },
1054
1092
  "ministral-3-14b": {
1055
1093
  // AA specific benchmarks
1056
- codingIndex: 10.9,
1094
+ codingIndex: 14.4,
1057
1095
  mathIndex: 30,
1058
1096
 
1059
1097
  // Academic benchmarks
@@ -1067,18 +1105,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
1067
1105
  supportsVision: false,
1068
1106
 
1069
1107
  // Metadata
1070
- lastUpdated: "2026-06-01",
1108
+ lastUpdated: "2026-07-01",
1071
1109
  originalModel: "Ministral 3 14B",
1072
1110
  },
1073
- "mistral-small-4-reasoning": {
1111
+ "devstral-small-2": {
1074
1112
  // AA specific benchmarks
1075
- codingIndex: 24.3,
1076
- mathIndex: undefined,
1113
+ codingIndex: 29.3,
1114
+ mathIndex: 34.3,
1077
1115
 
1078
1116
  // Academic benchmarks
1079
- mmluPro: undefined,
1080
- gpqa: 0.769,
1081
- hle: 0.095,
1117
+ mmluPro: 0.678,
1118
+ gpqa: 0.532,
1119
+ hle: 0.034,
1082
1120
 
1083
1121
  // Capabilities
1084
1122
  contextWindow: 8192,
@@ -1086,12 +1124,12 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
1086
1124
  supportsVision: false,
1087
1125
 
1088
1126
  // Metadata
1089
- lastUpdated: "2026-06-01",
1090
- originalModel: "Mistral Small 4 (Reasoning)",
1127
+ lastUpdated: "2026-07-01",
1128
+ originalModel: "Devstral Small 2",
1091
1129
  },
1092
1130
  "mistral-small-4-non-reasoning": {
1093
1131
  // AA specific benchmarks
1094
- codingIndex: 16.4,
1132
+ codingIndex: undefined,
1095
1133
  mathIndex: undefined,
1096
1134
 
1097
1135
  // Academic benchmarks
@@ -1105,18 +1143,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
1105
1143
  supportsVision: false,
1106
1144
 
1107
1145
  // Metadata
1108
- lastUpdated: "2026-06-01",
1146
+ lastUpdated: "2026-07-01",
1109
1147
  originalModel: "Mistral Small 4 (Non-reasoning)",
1110
1148
  },
1111
- "devstral-2": {
1149
+ "magistral-small-1.2": {
1112
1150
  // AA specific benchmarks
1113
- codingIndex: 23.7,
1114
- mathIndex: 36.7,
1151
+ codingIndex: 14.7,
1152
+ mathIndex: 80.3,
1115
1153
 
1116
1154
  // Academic benchmarks
1117
- mmluPro: 0.762,
1118
- gpqa: 0.594,
1119
- hle: 0.036,
1155
+ mmluPro: 0.768,
1156
+ gpqa: 0.663,
1157
+ hle: 0.061,
1120
1158
 
1121
1159
  // Capabilities
1122
1160
  contextWindow: 8192,
@@ -1124,8 +1162,8 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
1124
1162
  supportsVision: false,
1125
1163
 
1126
1164
  // Metadata
1127
- lastUpdated: "2026-06-01",
1128
- originalModel: "Devstral 2",
1165
+ lastUpdated: "2026-07-01",
1166
+ originalModel: "Magistral Small 1.2",
1129
1167
  },
1130
1168
  "ministral-3-3b": {
1131
1169
  // AA specific benchmarks
@@ -1143,18 +1181,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
1143
1181
  supportsVision: false,
1144
1182
 
1145
1183
  // Metadata
1146
- lastUpdated: "2026-06-01",
1184
+ lastUpdated: "2026-07-01",
1147
1185
  originalModel: "Ministral 3 3B",
1148
1186
  },
1149
- "deepseek-v4-pro-non-reasoning": {
1187
+ "ministral-3-8b": {
1150
1188
  // AA specific benchmarks
1151
- codingIndex: 38.4,
1152
- mathIndex: undefined,
1189
+ codingIndex: 9.7,
1190
+ mathIndex: 31.7,
1153
1191
 
1154
1192
  // Academic benchmarks
1155
- mmluPro: undefined,
1156
- gpqa: 0.717,
1157
- hle: 0.077,
1193
+ mmluPro: 0.642,
1194
+ gpqa: 0.471,
1195
+ hle: 0.043,
1158
1196
 
1159
1197
  // Capabilities
1160
1198
  contextWindow: 8192,
@@ -1162,18 +1200,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
1162
1200
  supportsVision: false,
1163
1201
 
1164
1202
  // Metadata
1165
- lastUpdated: "2026-06-01",
1166
- originalModel: "DeepSeek V4 Pro (Non-reasoning)",
1203
+ lastUpdated: "2026-07-01",
1204
+ originalModel: "Ministral 3 8B",
1167
1205
  },
1168
- "deepseek-v4-pro-reasoning-high-effort": {
1206
+ "devstral-2": {
1169
1207
  // AA specific benchmarks
1170
- codingIndex: 43.2,
1171
- mathIndex: undefined,
1208
+ codingIndex: 31.3,
1209
+ mathIndex: 36.7,
1172
1210
 
1173
1211
  // Academic benchmarks
1174
- mmluPro: undefined,
1175
- gpqa: 0.905,
1176
- hle: 0.335,
1212
+ mmluPro: 0.762,
1213
+ gpqa: 0.594,
1214
+ hle: 0.036,
1177
1215
 
1178
1216
  // Capabilities
1179
1217
  contextWindow: 8192,
@@ -1181,18 +1219,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
1181
1219
  supportsVision: false,
1182
1220
 
1183
1221
  // Metadata
1184
- lastUpdated: "2026-06-01",
1185
- originalModel: "DeepSeek V4 Pro (Reasoning, High Effort)",
1222
+ lastUpdated: "2026-07-01",
1223
+ originalModel: "Devstral 2",
1186
1224
  },
1187
- "deepseek-v4-pro-reasoning-max-effort": {
1225
+ "magistral-medium-1.2": {
1188
1226
  // AA specific benchmarks
1189
- codingIndex: 47.5,
1190
- mathIndex: undefined,
1227
+ codingIndex: 21.3,
1228
+ mathIndex: 82,
1191
1229
 
1192
1230
  // Academic benchmarks
1193
- mmluPro: undefined,
1194
- gpqa: 0.888,
1195
- hle: 0.359,
1231
+ mmluPro: 0.815,
1232
+ gpqa: 0.739,
1233
+ hle: 0.096,
1196
1234
 
1197
1235
  // Capabilities
1198
1236
  contextWindow: 8192,
@@ -1200,18 +1238,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
1200
1238
  supportsVision: false,
1201
1239
 
1202
1240
  // Metadata
1203
- lastUpdated: "2026-06-01",
1204
- originalModel: "DeepSeek V4 Pro (Reasoning, Max Effort)",
1241
+ lastUpdated: "2026-07-01",
1242
+ originalModel: "Magistral Medium 1.2",
1205
1243
  },
1206
- "deepseek-v4-flash-reasoning-max-effort": {
1244
+ "deepseek-v4-pro-reasoning-high-effort": {
1207
1245
  // AA specific benchmarks
1208
- codingIndex: 38.7,
1246
+ codingIndex: undefined,
1209
1247
  mathIndex: undefined,
1210
1248
 
1211
1249
  // Academic benchmarks
1212
1250
  mmluPro: undefined,
1213
- gpqa: 0.894,
1214
- hle: 0.321,
1251
+ gpqa: 0.905,
1252
+ hle: 0.335,
1215
1253
 
1216
1254
  // Capabilities
1217
1255
  contextWindow: 8192,
@@ -1219,18 +1257,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
1219
1257
  supportsVision: false,
1220
1258
 
1221
1259
  // Metadata
1222
- lastUpdated: "2026-06-01",
1223
- originalModel: "DeepSeek V4 Flash (Reasoning, Max Effort)",
1260
+ lastUpdated: "2026-07-01",
1261
+ originalModel: "DeepSeek V4 Pro (Reasoning, High Effort)",
1224
1262
  },
1225
- "deepseek-v4-flash-reasoning-high-effort": {
1263
+ "deepseek-v4-pro-reasoning-max-effort": {
1226
1264
  // AA specific benchmarks
1227
- codingIndex: 39.8,
1265
+ codingIndex: 59.4,
1228
1266
  mathIndex: undefined,
1229
1267
 
1230
1268
  // Academic benchmarks
1231
1269
  mmluPro: undefined,
1232
- gpqa: 0.867,
1233
- hle: 0.278,
1270
+ gpqa: 0.888,
1271
+ hle: 0.359,
1234
1272
 
1235
1273
  // Capabilities
1236
1274
  contextWindow: 8192,
@@ -1238,12 +1276,12 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
1238
1276
  supportsVision: false,
1239
1277
 
1240
1278
  // Metadata
1241
- lastUpdated: "2026-06-01",
1242
- originalModel: "DeepSeek V4 Flash (Reasoning, High Effort)",
1279
+ lastUpdated: "2026-07-01",
1280
+ originalModel: "DeepSeek V4 Pro (Reasoning, Max Effort)",
1243
1281
  },
1244
1282
  "deepseek-v4-flash-non-reasoning": {
1245
1283
  // AA specific benchmarks
1246
- codingIndex: 35.2,
1284
+ codingIndex: undefined,
1247
1285
  mathIndex: undefined,
1248
1286
 
1249
1287
  // Academic benchmarks
@@ -1257,37 +1295,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
1257
1295
  supportsVision: false,
1258
1296
 
1259
1297
  // Metadata
1260
- lastUpdated: "2026-06-01",
1298
+ lastUpdated: "2026-07-01",
1261
1299
  originalModel: "DeepSeek V4 Flash (Non-reasoning)",
1262
1300
  },
1263
- "r1-1776": {
1301
+ "deepseek-v4-flash-reasoning-high-effort": {
1264
1302
  // AA specific benchmarks
1265
1303
  codingIndex: undefined,
1266
1304
  mathIndex: undefined,
1267
1305
 
1268
1306
  // Academic benchmarks
1269
1307
  mmluPro: undefined,
1270
- gpqa: undefined,
1271
- hle: undefined,
1272
-
1273
- // Capabilities
1274
- contextWindow: 8192,
1275
- supportsReasoning: false,
1276
- supportsVision: false,
1277
-
1278
- // Metadata
1279
- lastUpdated: "2026-06-01",
1280
- originalModel: "R1 1776",
1281
- },
1282
- "falcon-h1r-7b": {
1283
- // AA specific benchmarks
1284
- codingIndex: 9.8,
1285
- mathIndex: 80,
1286
-
1287
- // Academic benchmarks
1288
- mmluPro: 0.725,
1289
- gpqa: 0.661,
1290
- hle: 0.108,
1308
+ gpqa: 0.867,
1309
+ hle: 0.278,
1291
1310
 
1292
1311
  // Capabilities
1293
1312
  contextWindow: 8192,
@@ -1295,18 +1314,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
1295
1314
  supportsVision: false,
1296
1315
 
1297
1316
  // Metadata
1298
- lastUpdated: "2026-06-01",
1299
- originalModel: "Falcon-H1R-7B",
1317
+ lastUpdated: "2026-07-01",
1318
+ originalModel: "DeepSeek V4 Flash (Reasoning, High Effort)",
1300
1319
  },
1301
- "grok-4.3-medium": {
1320
+ "deepseek-v4-flash-reasoning-max-effort": {
1302
1321
  // AA specific benchmarks
1303
- codingIndex: 35.1,
1322
+ codingIndex: 56.2,
1304
1323
  mathIndex: undefined,
1305
1324
 
1306
1325
  // Academic benchmarks
1307
1326
  mmluPro: undefined,
1308
- gpqa: 0.89,
1309
- hle: 0.281,
1327
+ gpqa: 0.894,
1328
+ hle: 0.321,
1310
1329
 
1311
1330
  // Capabilities
1312
1331
  contextWindow: 8192,
@@ -1314,18 +1333,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
1314
1333
  supportsVision: false,
1315
1334
 
1316
1335
  // Metadata
1317
- lastUpdated: "2026-06-01",
1318
- originalModel: "Grok 4.3 (medium)",
1336
+ lastUpdated: "2026-07-01",
1337
+ originalModel: "DeepSeek V4 Flash (Reasoning, Max Effort)",
1319
1338
  },
1320
- "grok-4.3-low": {
1339
+ "deepseek-v4-pro-non-reasoning": {
1321
1340
  // AA specific benchmarks
1322
- codingIndex: 31.6,
1341
+ codingIndex: undefined,
1323
1342
  mathIndex: undefined,
1324
1343
 
1325
1344
  // Academic benchmarks
1326
1345
  mmluPro: undefined,
1327
- gpqa: 0.843,
1328
- hle: 0.173,
1346
+ gpqa: 0.717,
1347
+ hle: 0.077,
1329
1348
 
1330
1349
  // Capabilities
1331
1350
  contextWindow: 8192,
@@ -1333,18 +1352,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
1333
1352
  supportsVision: false,
1334
1353
 
1335
1354
  // Metadata
1336
- lastUpdated: "2026-06-01",
1337
- originalModel: "Grok 4.3 (low)",
1355
+ lastUpdated: "2026-07-01",
1356
+ originalModel: "DeepSeek V4 Pro (Non-reasoning)",
1338
1357
  },
1339
- "grok-4.3-high": {
1358
+ "r1-1776": {
1340
1359
  // AA specific benchmarks
1341
- codingIndex: 41,
1360
+ codingIndex: undefined,
1342
1361
  mathIndex: undefined,
1343
1362
 
1344
1363
  // Academic benchmarks
1345
1364
  mmluPro: undefined,
1346
- gpqa: 0.901,
1347
- hle: 0.35,
1365
+ gpqa: undefined,
1366
+ hle: undefined,
1348
1367
 
1349
1368
  // Capabilities
1350
1369
  contextWindow: 8192,
@@ -1352,18 +1371,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
1352
1371
  supportsVision: false,
1353
1372
 
1354
1373
  // Metadata
1355
- lastUpdated: "2026-06-01",
1356
- originalModel: "Grok 4.3 (high)",
1374
+ lastUpdated: "2026-07-01",
1375
+ originalModel: "R1 1776",
1357
1376
  },
1358
- "grok-4.3-non-reasoning": {
1377
+ "falcon-h1r-7b": {
1359
1378
  // AA specific benchmarks
1360
- codingIndex: 25.1,
1361
- mathIndex: undefined,
1379
+ codingIndex: undefined,
1380
+ mathIndex: 80,
1362
1381
 
1363
1382
  // Academic benchmarks
1364
- mmluPro: undefined,
1365
- gpqa: 0.658,
1366
- hle: 0.065,
1383
+ mmluPro: 0.725,
1384
+ gpqa: 0.661,
1385
+ hle: 0.108,
1367
1386
 
1368
1387
  // Capabilities
1369
1388
  contextWindow: 8192,
@@ -1371,18 +1390,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
1371
1390
  supportsVision: false,
1372
1391
 
1373
1392
  // Metadata
1374
- lastUpdated: "2026-06-01",
1375
- originalModel: "Grok 4.3 (Non-reasoning)",
1393
+ lastUpdated: "2026-07-01",
1394
+ originalModel: "Falcon-H1R-7B",
1376
1395
  },
1377
- "nova-micro": {
1396
+ "grok-4.3-high": {
1378
1397
  // AA specific benchmarks
1379
- codingIndex: 4.1,
1380
- mathIndex: 6,
1398
+ codingIndex: 42.2,
1399
+ mathIndex: undefined,
1381
1400
 
1382
1401
  // Academic benchmarks
1383
- mmluPro: 0.531,
1384
- gpqa: 0.358,
1385
- hle: 0.047,
1402
+ mmluPro: undefined,
1403
+ gpqa: 0.901,
1404
+ hle: 0.35,
1386
1405
 
1387
1406
  // Capabilities
1388
1407
  contextWindow: 8192,
@@ -1390,18 +1409,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
1390
1409
  supportsVision: false,
1391
1410
 
1392
1411
  // Metadata
1393
- lastUpdated: "2026-06-01",
1394
- originalModel: "Nova Micro",
1412
+ lastUpdated: "2026-07-01",
1413
+ originalModel: "Grok 4.3 (high)",
1395
1414
  },
1396
- "nova-2.0-omni-low": {
1415
+ "grok-4.3-medium": {
1397
1416
  // AA specific benchmarks
1398
- codingIndex: 13.9,
1399
- mathIndex: 56,
1417
+ codingIndex: undefined,
1418
+ mathIndex: undefined,
1400
1419
 
1401
1420
  // Academic benchmarks
1402
- mmluPro: 0.798,
1403
- gpqa: 0.699,
1404
- hle: 0.04,
1421
+ mmluPro: undefined,
1422
+ gpqa: 0.89,
1423
+ hle: 0.281,
1405
1424
 
1406
1425
  // Capabilities
1407
1426
  contextWindow: 8192,
@@ -1409,18 +1428,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
1409
1428
  supportsVision: false,
1410
1429
 
1411
1430
  // Metadata
1412
- lastUpdated: "2026-06-01",
1413
- originalModel: "Nova 2.0 Omni (low)",
1431
+ lastUpdated: "2026-07-01",
1432
+ originalModel: "Grok 4.3 (medium)",
1414
1433
  },
1415
- "nova-2.0-pro-preview-medium": {
1434
+ "grok-4.3-low": {
1416
1435
  // AA specific benchmarks
1417
- codingIndex: 30.4,
1418
- mathIndex: 89,
1436
+ codingIndex: undefined,
1437
+ mathIndex: undefined,
1419
1438
 
1420
1439
  // Academic benchmarks
1421
- mmluPro: 0.83,
1422
- gpqa: 0.785,
1423
- hle: 0.089,
1440
+ mmluPro: undefined,
1441
+ gpqa: 0.843,
1442
+ hle: 0.173,
1424
1443
 
1425
1444
  // Capabilities
1426
1445
  contextWindow: 8192,
@@ -1428,18 +1447,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
1428
1447
  supportsVision: false,
1429
1448
 
1430
1449
  // Metadata
1431
- lastUpdated: "2026-06-01",
1432
- originalModel: "Nova 2.0 Pro Preview (medium)",
1450
+ lastUpdated: "2026-07-01",
1451
+ originalModel: "Grok 4.3 (low)",
1433
1452
  },
1434
- "nova-2.0-lite-medium": {
1453
+ "grok-4.3-non-reasoning": {
1435
1454
  // AA specific benchmarks
1436
- codingIndex: 23.9,
1437
- mathIndex: 88.7,
1455
+ codingIndex: 35.2,
1456
+ mathIndex: undefined,
1438
1457
 
1439
1458
  // Academic benchmarks
1440
- mmluPro: 0.813,
1441
- gpqa: 0.768,
1442
- hle: 0.086,
1459
+ mmluPro: undefined,
1460
+ gpqa: 0.658,
1461
+ hle: 0.065,
1443
1462
 
1444
1463
  // Capabilities
1445
1464
  contextWindow: 8192,
@@ -1447,18 +1466,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
1447
1466
  supportsVision: false,
1448
1467
 
1449
1468
  // Metadata
1450
- lastUpdated: "2026-06-01",
1451
- originalModel: "Nova 2.0 Lite (medium)",
1469
+ lastUpdated: "2026-07-01",
1470
+ originalModel: "Grok 4.3 (Non-reasoning)",
1452
1471
  },
1453
- "nova-2.0-lite-high": {
1472
+ "grok-build-0.1-0616": {
1454
1473
  // AA specific benchmarks
1455
- codingIndex: 23.4,
1456
- mathIndex: 94.3,
1474
+ codingIndex: 51.5,
1475
+ mathIndex: undefined,
1457
1476
 
1458
1477
  // Academic benchmarks
1459
- mmluPro: 0.818,
1460
- gpqa: 0.811,
1461
- hle: 0.109,
1478
+ mmluPro: undefined,
1479
+ gpqa: 0.895,
1480
+ hle: 0.36,
1462
1481
 
1463
1482
  // Capabilities
1464
1483
  contextWindow: 8192,
@@ -1466,18 +1485,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
1466
1485
  supportsVision: false,
1467
1486
 
1468
1487
  // Metadata
1469
- lastUpdated: "2026-06-01",
1470
- originalModel: "Nova 2.0 Lite (high)",
1488
+ lastUpdated: "2026-07-01",
1489
+ originalModel: "Grok Build 0.1 0616",
1471
1490
  },
1472
- "nova-2.0-pro-preview-non-reasoning": {
1491
+ "nova-micro": {
1473
1492
  // AA specific benchmarks
1474
- codingIndex: 20.5,
1475
- mathIndex: 30.7,
1493
+ codingIndex: undefined,
1494
+ mathIndex: 6,
1476
1495
 
1477
1496
  // Academic benchmarks
1478
- mmluPro: 0.772,
1479
- gpqa: 0.636,
1480
- hle: 0.04,
1497
+ mmluPro: 0.531,
1498
+ gpqa: 0.358,
1499
+ hle: 0.047,
1481
1500
 
1482
1501
  // Capabilities
1483
1502
  contextWindow: 8192,
@@ -1485,18 +1504,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
1485
1504
  supportsVision: false,
1486
1505
 
1487
1506
  // Metadata
1488
- lastUpdated: "2026-06-01",
1489
- originalModel: "Nova 2.0 Pro Preview (Non-reasoning)",
1507
+ lastUpdated: "2026-07-01",
1508
+ originalModel: "Nova Micro",
1490
1509
  },
1491
- "nova-2.0-lite-low": {
1510
+ "nova-2.0-pro-preview-low": {
1492
1511
  // AA specific benchmarks
1493
- codingIndex: 13.6,
1494
- mathIndex: 46.7,
1512
+ codingIndex: 25.9,
1513
+ mathIndex: 63.3,
1495
1514
 
1496
1515
  // Academic benchmarks
1497
- mmluPro: 0.788,
1498
- gpqa: 0.698,
1499
- hle: 0.042,
1516
+ mmluPro: 0.822,
1517
+ gpqa: 0.751,
1518
+ hle: 0.052,
1500
1519
 
1501
1520
  // Capabilities
1502
1521
  contextWindow: 8192,
@@ -1504,18 +1523,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
1504
1523
  supportsVision: false,
1505
1524
 
1506
1525
  // Metadata
1507
- lastUpdated: "2026-06-01",
1508
- originalModel: "Nova 2.0 Lite (low)",
1526
+ lastUpdated: "2026-07-01",
1527
+ originalModel: "Nova 2.0 Pro Preview (low)",
1509
1528
  },
1510
- "nova-2.0-lite-non-reasoning": {
1529
+ "nova-2.0-omni-medium": {
1511
1530
  // AA specific benchmarks
1512
- codingIndex: 12.5,
1513
- mathIndex: 33.7,
1531
+ codingIndex: undefined,
1532
+ mathIndex: 89.7,
1514
1533
 
1515
1534
  // Academic benchmarks
1516
- mmluPro: 0.743,
1517
- gpqa: 0.603,
1518
- hle: 0.03,
1535
+ mmluPro: 0.809,
1536
+ gpqa: 0.76,
1537
+ hle: 0.068,
1519
1538
 
1520
1539
  // Capabilities
1521
1540
  contextWindow: 8192,
@@ -1523,18 +1542,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
1523
1542
  supportsVision: false,
1524
1543
 
1525
1544
  // Metadata
1526
- lastUpdated: "2026-06-01",
1527
- originalModel: "Nova 2.0 Lite (Non-reasoning)",
1545
+ lastUpdated: "2026-07-01",
1546
+ originalModel: "Nova 2.0 Omni (medium)",
1528
1547
  },
1529
- "nova-premier": {
1548
+ "nova-2.0-pro-preview-medium": {
1530
1549
  // AA specific benchmarks
1531
- codingIndex: 13.8,
1532
- mathIndex: 17.3,
1550
+ codingIndex: 34,
1551
+ mathIndex: 89,
1533
1552
 
1534
1553
  // Academic benchmarks
1535
- mmluPro: 0.733,
1536
- gpqa: 0.569,
1537
- hle: 0.047,
1554
+ mmluPro: 0.83,
1555
+ gpqa: 0.785,
1556
+ hle: 0.089,
1538
1557
 
1539
1558
  // Capabilities
1540
1559
  contextWindow: 8192,
@@ -1542,12 +1561,12 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
1542
1561
  supportsVision: false,
1543
1562
 
1544
1563
  // Metadata
1545
- lastUpdated: "2026-06-01",
1546
- originalModel: "Nova Premier",
1564
+ lastUpdated: "2026-07-01",
1565
+ originalModel: "Nova 2.0 Pro Preview (medium)",
1547
1566
  },
1548
1567
  "nova-2.0-omni-non-reasoning": {
1549
1568
  // AA specific benchmarks
1550
- codingIndex: 13.8,
1569
+ codingIndex: undefined,
1551
1570
  mathIndex: 37,
1552
1571
 
1553
1572
  // Academic benchmarks
@@ -1561,18 +1580,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
1561
1580
  supportsVision: false,
1562
1581
 
1563
1582
  // Metadata
1564
- lastUpdated: "2026-06-01",
1583
+ lastUpdated: "2026-07-01",
1565
1584
  originalModel: "Nova 2.0 Omni (Non-reasoning)",
1566
1585
  },
1567
- "nova-2.0-omni-medium": {
1586
+ "nova-2.0-lite-high": {
1568
1587
  // AA specific benchmarks
1569
- codingIndex: 15.1,
1570
- mathIndex: 89.7,
1588
+ codingIndex: 23,
1589
+ mathIndex: 94.3,
1571
1590
 
1572
1591
  // Academic benchmarks
1573
- mmluPro: 0.809,
1574
- gpqa: 0.76,
1575
- hle: 0.068,
1592
+ mmluPro: 0.818,
1593
+ gpqa: 0.811,
1594
+ hle: 0.109,
1576
1595
 
1577
1596
  // Capabilities
1578
1597
  contextWindow: 8192,
@@ -1580,18 +1599,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
1580
1599
  supportsVision: false,
1581
1600
 
1582
1601
  // Metadata
1583
- lastUpdated: "2026-06-01",
1584
- originalModel: "Nova 2.0 Omni (medium)",
1602
+ lastUpdated: "2026-07-01",
1603
+ originalModel: "Nova 2.0 Lite (high)",
1585
1604
  },
1586
- "nova-2.0-pro-preview-low": {
1605
+ "nova-2.0-lite-low": {
1587
1606
  // AA specific benchmarks
1588
- codingIndex: 24.5,
1589
- mathIndex: 63.3,
1607
+ codingIndex: undefined,
1608
+ mathIndex: 46.7,
1590
1609
 
1591
1610
  // Academic benchmarks
1592
- mmluPro: 0.822,
1593
- gpqa: 0.751,
1594
- hle: 0.052,
1611
+ mmluPro: 0.788,
1612
+ gpqa: 0.698,
1613
+ hle: 0.042,
1595
1614
 
1596
1615
  // Capabilities
1597
1616
  contextWindow: 8192,
@@ -1599,18 +1618,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
1599
1618
  supportsVision: false,
1600
1619
 
1601
1620
  // Metadata
1602
- lastUpdated: "2026-06-01",
1603
- originalModel: "Nova 2.0 Pro Preview (low)",
1621
+ lastUpdated: "2026-07-01",
1622
+ originalModel: "Nova 2.0 Lite (low)",
1604
1623
  },
1605
- "phi-4": {
1624
+ "nova-2.0-lite-non-reasoning": {
1606
1625
  // AA specific benchmarks
1607
- codingIndex: 11.2,
1608
- mathIndex: 18,
1626
+ codingIndex: undefined,
1627
+ mathIndex: 33.7,
1609
1628
 
1610
1629
  // Academic benchmarks
1611
- mmluPro: 0.714,
1612
- gpqa: 0.575,
1613
- hle: 0.041,
1630
+ mmluPro: 0.743,
1631
+ gpqa: 0.603,
1632
+ hle: 0.03,
1614
1633
 
1615
1634
  // Capabilities
1616
1635
  contextWindow: 8192,
@@ -1618,18 +1637,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
1618
1637
  supportsVision: false,
1619
1638
 
1620
1639
  // Metadata
1621
- lastUpdated: "2026-06-01",
1622
- originalModel: "Phi-4",
1640
+ lastUpdated: "2026-07-01",
1641
+ originalModel: "Nova 2.0 Lite (Non-reasoning)",
1623
1642
  },
1624
- "phi-4-mini-instruct": {
1643
+ "nova-2.0-pro-preview-non-reasoning": {
1625
1644
  // AA specific benchmarks
1626
- codingIndex: 3.6,
1627
- mathIndex: 6.7,
1645
+ codingIndex: 20.9,
1646
+ mathIndex: 30.7,
1628
1647
 
1629
1648
  // Academic benchmarks
1630
- mmluPro: 0.465,
1631
- gpqa: 0.331,
1632
- hle: 0.042,
1649
+ mmluPro: 0.772,
1650
+ gpqa: 0.636,
1651
+ hle: 0.04,
1633
1652
 
1634
1653
  // Capabilities
1635
1654
  contextWindow: 8192,
@@ -1637,37 +1656,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
1637
1656
  supportsVision: false,
1638
1657
 
1639
1658
  // Metadata
1640
- lastUpdated: "2026-06-01",
1641
- originalModel: "Phi-4 Mini Instruct",
1659
+ lastUpdated: "2026-07-01",
1660
+ originalModel: "Nova 2.0 Pro Preview (Non-reasoning)",
1642
1661
  },
1643
- "phi-4-multimodal-instruct": {
1662
+ "nova-2.0-omni-low": {
1644
1663
  // AA specific benchmarks
1645
1664
  codingIndex: undefined,
1646
- mathIndex: undefined,
1647
-
1648
- // Academic benchmarks
1649
- mmluPro: 0.485,
1650
- gpqa: 0.315,
1651
- hle: 0.044,
1652
-
1653
- // Capabilities
1654
- contextWindow: 8192,
1655
- supportsReasoning: false,
1656
- supportsVision: false,
1657
-
1658
- // Metadata
1659
- lastUpdated: "2026-06-01",
1660
- originalModel: "Phi-4 Multimodal Instruct",
1661
- },
1662
- "lfm2.5-vl-1.6b": {
1663
- // AA specific benchmarks
1664
- codingIndex: 1,
1665
- mathIndex: undefined,
1665
+ mathIndex: 56,
1666
1666
 
1667
1667
  // Academic benchmarks
1668
- mmluPro: undefined,
1669
- gpqa: 0.289,
1670
- hle: 0.051,
1668
+ mmluPro: 0.798,
1669
+ gpqa: 0.699,
1670
+ hle: 0.04,
1671
1671
 
1672
1672
  // Capabilities
1673
1673
  contextWindow: 8192,
@@ -1675,18 +1675,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
1675
1675
  supportsVision: false,
1676
1676
 
1677
1677
  // Metadata
1678
- lastUpdated: "2026-06-01",
1679
- originalModel: "LFM2.5-VL-1.6B",
1678
+ lastUpdated: "2026-07-01",
1679
+ originalModel: "Nova 2.0 Omni (low)",
1680
1680
  },
1681
- "lfm2-24b-a2b": {
1681
+ "nova-premier": {
1682
1682
  // AA specific benchmarks
1683
- codingIndex: 3.6,
1684
- mathIndex: undefined,
1683
+ codingIndex: undefined,
1684
+ mathIndex: 17.3,
1685
1685
 
1686
1686
  // Academic benchmarks
1687
- mmluPro: undefined,
1688
- gpqa: 0.474,
1689
- hle: 0.044,
1687
+ mmluPro: 0.733,
1688
+ gpqa: 0.569,
1689
+ hle: 0.047,
1690
1690
 
1691
1691
  // Capabilities
1692
1692
  contextWindow: 8192,
@@ -1694,18 +1694,18 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
1694
1694
  supportsVision: false,
1695
1695
 
1696
1696
  // Metadata
1697
- lastUpdated: "2026-06-01",
1698
- originalModel: "LFM2 24B A2B",
1697
+ lastUpdated: "2026-07-01",
1698
+ originalModel: "Nova Premier",
1699
1699
  },
1700
- "lfm2-8b-a1b": {
1700
+ "nova-2.0-lite-medium": {
1701
1701
  // AA specific benchmarks
1702
- codingIndex: 2.3,
1703
- mathIndex: 25.3,
1702
+ codingIndex: undefined,
1703
+ mathIndex: 88.7,
1704
1704
 
1705
1705
  // Academic benchmarks
1706
- mmluPro: 0.505,
1707
- gpqa: 0.344,
1708
- hle: 0.049,
1706
+ mmluPro: 0.813,
1707
+ gpqa: 0.768,
1708
+ hle: 0.086,
1709
1709
 
1710
1710
  // Capabilities
1711
1711
  contextWindow: 8192,
@@ -1713,7 +1713,7 @@ export const BENCHMARKS_CHUNK_0: Record<string, HardcodedBenchmark> = {
1713
1713
  supportsVision: false,
1714
1714
 
1715
1715
  // Metadata
1716
- lastUpdated: "2026-06-01",
1717
- originalModel: "LFM2 8B A1B",
1716
+ lastUpdated: "2026-07-01",
1717
+ originalModel: "Nova 2.0 Lite (medium)",
1718
1718
  },
1719
1719
  };