Benchmark
IFBench
Instruction following
- Categories
- instruction-following
- Openness
- unknown
- Source
- artificial_analysis
- Reported scores
- 450
Reported scores
artificial_analysis
| Model | Organization | Reported value | Reported | Evidence |
|---|---|---|---|---|
| Apertus 70B Instruct | Swiss AI Initiative | 0.258503401360544 | 2025-09-02 | evidence |
| Apertus 8B Instruct | Swiss AI Initiative | 0.223809523809524 | 2025-09-02 | evidence |
| Apriel-v1.5-15B-Thinker | ServiceNow | 0.617006802721088 | 2025-09-30 | evidence |
| Apriel-v1.6-15B-Thinker | ServiceNow | 0.691156462585034 | 2025-11-25 | evidence |
| Claude 3 Haiku | Anthropic | 0.361224489795918 | 2024-03-04 | evidence |
| Claude 3.5 Haiku | Anthropic | 0.427891156462585 | 2024-10-22 | evidence |
| Claude 3.7 Sonnet (Non-reasoning) | Anthropic | 0.440136054421769 | 2025-02-24 | evidence |
| Claude 3.7 Sonnet (Reasoning) | Anthropic | 0.482993197278912 | 2025-02-24 | evidence |
| Claude 4 Opus (Non-reasoning) | Anthropic | 0.43265306122449 | 2025-05-22 | evidence |
| Claude 4 Opus (Reasoning) | Anthropic | 0.537414965986395 | 2025-05-22 | evidence |
| Claude 4 Sonnet (Non-reasoning) | Anthropic | 0.453741496598639 | 2025-05-22 | evidence |
| Claude 4 Sonnet (Reasoning) | Anthropic | 0.546938775510204 | 2025-05-22 | evidence |
| Claude 4.1 Opus (Reasoning) | Anthropic | 0.554421768707483 | 2025-08-05 | evidence |
| Claude 4.5 Haiku (Non-reasoning) | Anthropic | 0.420408163265306 | 2025-10-15 | evidence |
| Claude 4.5 Haiku (Reasoning) | Anthropic | 0.542857142857143 | 2025-10-15 | evidence |
| Claude 4.5 Sonnet (Non-reasoning) | Anthropic | 0.426530612244898 | 2025-09-29 | evidence |
| Claude 4.5 Sonnet (Reasoning) | Anthropic | 0.572789115646258 | 2025-09-29 | evidence |
| Claude Fable 5 (Adaptive Reasoning, Max Effort, Opus 4.8 Fallback) | Anthropic | 0.63469387755102 | 2026-06-09 | evidence |
| Claude Opus 4.5 (Non-reasoning) | Anthropic | 0.429931972789116 | 2025-11-24 | evidence |
| Claude Opus 4.5 (Reasoning) | Anthropic | 0.579591836734694 | 2025-11-24 | evidence |
| Claude Opus 4.6 (Adaptive Reasoning, Max Effort) | Anthropic | 0.531292517006803 | 2026-02-05 | evidence |
| Claude Opus 4.6 (Non-reasoning, High Effort) | Anthropic | 0.446258503401361 | 2026-02-05 | evidence |
| Claude Opus 4.7 (Adaptive Reasoning, Max Effort) | Anthropic | 0.586394557823129 | 2026-04-16 | evidence |
| Claude Opus 4.7 (Non-reasoning, High Effort) | Anthropic | 0.436054421768707 | 2026-04-16 | evidence |
| Claude Opus 4.8 (Adaptive Reasoning, Max Effort) | Anthropic | 0.622448979591837 | 2026-05-28 | evidence |
| Claude Sonnet 4.6 (Adaptive Reasoning, Max Effort) | Anthropic | 0.565986394557823 | 2026-02-17 | evidence |
| Claude Sonnet 4.6 (Non-reasoning, High Effort) | Anthropic | 0.41156462585034 | 2026-02-17 | evidence |
| Claude Sonnet 4.6 (Non-reasoning, Low Effort) | Anthropic | 0.423809523809524 | 2026-02-17 | evidence |
| Cogito v2.1 (Reasoning) | Deep Cogito | 0.462585034013605 | 2025-11-18 | evidence |
| Command A | Cohere | 0.364625850340136 | 2025-03-13 | evidence |
| Command A+ | Cohere | 0.739455782312925 | 2026-05-20 | evidence |
| DeepSeek R1 (Jan '25) | DeepSeek | 0.389795918367347 | 2025-01-20 | evidence |
| DeepSeek R1 0528 (May '25) | DeepSeek | 0.395918367346939 | 2025-05-28 | evidence |
| DeepSeek R1 0528 Qwen3 8B | DeepSeek | 0.198639455782313 | 2025-05-29 | evidence |
| DeepSeek R1 Distill Llama 70B | DeepSeek | 0.275510204081633 | 2025-01-20 | evidence |
| DeepSeek R1 Distill Llama 8B | DeepSeek | 0.175510204081633 | 2025-01-20 | evidence |
| DeepSeek R1 Distill Qwen 1.5B | DeepSeek | 0.131972789115646 | 2025-01-20 | evidence |
| DeepSeek R1 Distill Qwen 14B | DeepSeek | 0.22108843537415 | 2025-01-20 | evidence |
| DeepSeek R1 Distill Qwen 32B | DeepSeek | 0.229251700680272 | 2025-01-20 | evidence |
| DeepSeek V3 (Dec '24) | DeepSeek | 0.347619047619048 | 2024-12-26 | evidence |
| DeepSeek V3 0324 | DeepSeek | 0.410204081632653 | 2025-03-25 | evidence |
| DeepSeek V3.1 (Non-reasoning) | DeepSeek | 0.378231292517007 | 2025-08-21 | evidence |
| DeepSeek V3.1 (Reasoning) | DeepSeek | 0.414965986394558 | 2025-08-21 | evidence |
| DeepSeek V3.1 Terminus (Non-reasoning) | DeepSeek | 0.412244897959184 | 2025-09-22 | evidence |
| DeepSeek V3.1 Terminus (Reasoning) | DeepSeek | 0.570068027210884 | 2025-09-22 | evidence |
| DeepSeek V3.2 (Non-reasoning) | DeepSeek | 0.490476190476191 | 2025-12-01 | evidence |
| DeepSeek V3.2 (Reasoning) | DeepSeek | 0.606802721088435 | 2025-12-01 | evidence |
| DeepSeek V3.2 Exp (Non-reasoning) | DeepSeek | 0.431292517006803 | 2025-09-29 | evidence |
| DeepSeek V3.2 Exp (Reasoning) | DeepSeek | 0.541496598639456 | 2025-09-29 | evidence |
| DeepSeek V3.2 Speciale | DeepSeek | 0.638775510204082 | 2025-12-01 | evidence |
| DeepSeek V4 Flash (Non-reasoning) | DeepSeek | 0.472108843537415 | 2026-04-24 | evidence |
| DeepSeek V4 Flash (Reasoning, High Effort) | DeepSeek | 0.734693877551021 | 2026-04-24 | evidence |
| DeepSeek V4 Flash (Reasoning, Max Effort) | DeepSeek | 0.791836734693877 | 2026-04-24 | evidence |
| DeepSeek V4 Pro (Non-reasoning) | DeepSeek | 0.457823129251701 | 2026-04-24 | evidence |
| DeepSeek V4 Pro (Reasoning, High Effort) | DeepSeek | 0.712925170068027 | 2026-04-24 | evidence |
| DeepSeek V4 Pro (Reasoning, Max Effort) | DeepSeek | 0.764625850340136 | 2026-04-24 | evidence |
| Devstral 2 | Mistral | 0.380952380952381 | 2025-12-09 | evidence |
| Devstral Medium | Mistral | 0.299319727891156 | 2025-07-10 | evidence |
| Devstral Small (Jul '25) | Mistral | 0.345578231292517 | 2025-07-10 | evidence |
| Devstral Small (May '25) | Mistral | 0.316326530612245 | 2025-05-21 | evidence |
| Devstral Small 2 | Mistral | 0.31156462585034 | 2025-12-09 | evidence |
| DiffusionGemma 26B A4B | 0.594557823129252 | 2026-06-10 | evidence | |
| Doubao Seed Code | ByteDance Seed | 0.514285714285714 | 2025-11-11 | evidence |
| ERNIE 4.5 300B A47B | Baidu | 0.391156462585034 | 2025-06-30 | evidence |
| ERNIE 5.0 Thinking Preview | Baidu | 0.413605442176871 | 2025-11-13 | evidence |
| Exaone 4.0 1.2B (Non-reasoning) | LG AI Research | 0.253061224489796 | 2025-07-15 | evidence |
| Exaone 4.0 1.2B (Reasoning) | LG AI Research | 0.229931972789116 | 2025-07-15 | evidence |
| EXAONE 4.0 32B (Non-reasoning) | LG AI Research | 0.33469387755102 | 2025-07-15 | evidence |
| EXAONE 4.0 32B (Reasoning) | LG AI Research | 0.363265306122449 | 2025-07-15 | evidence |
| EXAONE 4.5 33B | LG AI Research | 0.579591836734694 | 2026-04-09 | evidence |
| Falcon-H1R-7B | TII UAE | 0.54421768707483 | 2026-01-04 | evidence |
| Gemini 2.0 Flash (Feb '25) | 0.402040816326531 | 2025-02-05 | evidence | |
| Gemini 2.5 Flash (Non-reasoning) | 0.389795918367347 | 2025-05-20 | evidence | |
| Gemini 2.5 Flash (Reasoning) | 0.502721088435374 | 2025-05-20 | evidence | |
| Gemini 2.5 Flash Preview (Sep '25) (Non-reasoning) | 0.435374149659864 | 2025-09-25 | evidence | |
| Gemini 2.5 Flash Preview (Sep '25) (Reasoning) | 0.52312925170068 | 2025-09-25 | evidence | |
| Gemini 2.5 Flash-Lite (Non-reasoning) | 0.314965986394558 | 2025-06-17 | evidence | |
| Gemini 2.5 Flash-Lite (Reasoning) | 0.498639455782313 | 2025-06-17 | evidence | |
| Gemini 2.5 Flash-Lite Preview (Sep '25) (Non-reasoning) | 0.418367346938776 | 2025-09-25 | evidence | |
| Gemini 2.5 Flash-Lite Preview (Sep '25) (Reasoning) | 0.525850340136054 | 2025-09-08 | evidence | |
| Gemini 2.5 Pro | 0.487074829931973 | 2025-06-05 | evidence | |
| Gemini 3 Flash Preview (Non-reasoning) | 0.551020408163265 | 2025-12-17 | evidence | |
| Gemini 3 Flash Preview (Reasoning) | 0.779591836734694 | 2025-12-17 | evidence | |
| Gemini 3 Pro Preview (high) | 0.704081632653061 | 2025-11-18 | evidence | |
| Gemini 3 Pro Preview (low) | 0.496598639455782 | 2025-11-18 | evidence | |
| Gemini 3.1 Flash-Lite | 0.772108843537415 | 2026-03-03 | evidence | |
| Gemini 3.1 Pro Preview | 0.771428571428571 | 2026-02-19 | evidence | |
| Gemini 3.5 Flash (high) | 0.763265306122449 | 2026-05-19 | evidence | |
| Gemini 3.5 Flash (medium) | 0.745578231292517 | 2026-05-19 | evidence | |
| Gemini 3.5 Flash (minimal) | 0.472789115646259 | 2026-05-19 | evidence | |
| Gemma 3 12B Instruct | 0.36734693877551 | 2025-03-12 | evidence | |
| Gemma 3 1B Instruct | 0.199319727891156 | 2025-03-13 | evidence | |
| Gemma 3 270M | 0.12108843537415 | 2025-08-14 | evidence | |
| Gemma 3 27B Instruct | 0.318367346938776 | 2025-03-12 | evidence | |
| Gemma 3 4B Instruct | 0.282993197278912 | 2025-03-12 | evidence | |
| Gemma 3n E2B Instruct | 0.220408163265306 | 2025-06-26 | evidence | |
| Gemma 3n E4B Instruct | 0.27891156462585 | 2025-06-26 | evidence | |
| Gemma 4 12B (Non-reasoning) | 0.451700680272109 | 2026-06-03 | evidence | |
| Gemma 4 12B (Reasoning) | 0.735374149659864 | 2026-06-03 | evidence | |
| Gemma 4 26B A4B (Non-reasoning) | 0.454421768707483 | 2026-04-02 | evidence | |
| Gemma 4 26B A4B (Reasoning) | 0.724489795918367 | 2026-04-02 | evidence | |
| Gemma 4 31B (Non-reasoning) | 0.53469387755102 | 2026-04-02 | evidence | |
| Gemma 4 31B (Reasoning) | 0.75578231292517 | 2026-04-02 | evidence | |
| Gemma 4 E2B (Non-reasoning) | 0.336054421768707 | 2026-04-02 | evidence | |
| Gemma 4 E2B (Reasoning) | 0.380272108843537 | 2026-04-02 | evidence | |
| Gemma 4 E4B (Non-reasoning) | 0.405442176870748 | 2026-04-03 | evidence | |
| Gemma 4 E4B (Reasoning) | 0.442176870748299 | 2026-04-03 | evidence | |
| GLM 5V Turbo (Reasoning) | Z.ai | 0.610884353741497 | 2026-04-01 | evidence |
| GLM-4.5 (Reasoning) | Z.ai | 0.440816326530612 | 2025-07-28 | evidence |
| GLM-4.5-Air | Z.ai | 0.375510204081633 | 2025-07-28 | evidence |
| GLM-4.5V (Non-reasoning) | Z.ai | 0.286394557823129 | 2025-08-11 | evidence |
| GLM-4.5V (Reasoning) | Z.ai | 0.342176870748299 | 2025-08-11 | evidence |
| GLM-4.6 (Non-reasoning) | Z.ai | 0.36734693877551 | 2025-09-30 | evidence |
| GLM-4.6 (Reasoning) | Z.ai | 0.434013605442177 | 2025-09-30 | evidence |
| GLM-4.6V (Non-reasoning) | Z.ai | 0.27891156462585 | 2025-12-08 | evidence |
| GLM-4.6V (Reasoning) | Z.ai | 0.300680272108844 | 2025-12-08 | evidence |
| GLM-4.7 (Non-reasoning) | Z.ai | 0.546258503401361 | 2025-12-22 | evidence |
| GLM-4.7 (Reasoning) | Z.ai | 0.67891156462585 | 2025-12-22 | evidence |
| GLM-4.7-Flash (Non-reasoning) | Z.ai | 0.462585034013605 | 2026-01-19 | evidence |
| GLM-4.7-Flash (Reasoning) | Z.ai | 0.608163265306122 | 2026-01-19 | evidence |
| GLM-5 (Non-reasoning) | Z.ai | 0.551700680272109 | 2026-02-11 | evidence |
| GLM-5 (Reasoning) | Z.ai | 0.72312925170068 | 2026-02-11 | evidence |
| GLM-5-Turbo | Z.ai | 0.731972789115646 | 2026-03-15 | evidence |
| GLM-5.1 (Non-reasoning) | Z.ai | 0.519727891156463 | 2026-04-07 | evidence |
| GLM-5.1 (Reasoning) | Z.ai | 0.762585034013605 | 2026-04-07 | evidence |
| GLM-5.2 (max) | Z.ai | 0.733333333333333 | 2026-06-16 | evidence |
| GPT-4 | OpenAI | 0.331972789115646 | 2023-03-14 | evidence |
| GPT-4.1 | OpenAI | 0.429931972789116 | 2025-04-14 | evidence |
| GPT-4.1 mini | OpenAI | 0.382993197278912 | 2025-04-14 | evidence |
| GPT-4.1 nano | OpenAI | 0.320408163265306 | 2025-04-14 | evidence |
| GPT-4o (Aug '24) | OpenAI | 0.359863945578231 | 2024-08-06 | evidence |
| GPT-4o (Nov '24) | OpenAI | 0.342857142857143 | 2024-11-20 | evidence |
| GPT-4o mini | OpenAI | 0.30952380952381 | 2024-07-18 | evidence |
| GPT-5 (ChatGPT) | OpenAI | 0.450340136054422 | 2025-08-07 | evidence |
| GPT-5 (high) | OpenAI | 0.730612244897959 | 2025-08-07 | evidence |
| GPT-5 (low) | OpenAI | 0.665986394557823 | 2025-08-07 | evidence |
| GPT-5 (medium) | OpenAI | 0.706122448979592 | 2025-08-07 | evidence |
| GPT-5 (minimal) | OpenAI | 0.45578231292517 | 2025-08-07 | evidence |
| GPT-5 Codex (high) | OpenAI | 0.741496598639456 | 2025-09-23 | evidence |
| GPT-5 mini (high) | OpenAI | 0.754421768707483 | 2025-08-07 | evidence |
| GPT-5 mini (medium) | OpenAI | 0.71156462585034 | 2025-08-07 | evidence |
| GPT-5 mini (minimal) | OpenAI | 0.45578231292517 | 2025-08-07 | evidence |
| GPT-5 nano (high) | OpenAI | 0.675510204081633 | 2025-08-07 | evidence |
| GPT-5 nano (medium) | OpenAI | 0.659183673469388 | 2025-08-07 | evidence |
| GPT-5 nano (minimal) | OpenAI | 0.325170068027211 | 2025-08-07 | evidence |
| GPT-5.1 (high) | OpenAI | 0.728571428571429 | 2025-11-13 | evidence |
| GPT-5.1 (Non-reasoning) | OpenAI | 0.431972789115646 | 2025-11-13 | evidence |
| GPT-5.1 Codex (high) | OpenAI | 0.7 | 2025-11-13 | evidence |
| GPT-5.1 Codex mini (high) | OpenAI | 0.67891156462585 | 2025-11-13 | evidence |
| GPT-5.2 (medium) | OpenAI | 0.652380952380953 | 2025-12-11 | evidence |
| GPT-5.2 (Non-reasoning) | OpenAI | 0.474149659863946 | 2025-12-11 | evidence |
| GPT-5.2 (xhigh) | OpenAI | 0.754421768707483 | 2025-12-11 | evidence |
| GPT-5.2 Codex (xhigh) | OpenAI | 0.776190476190476 | 2025-12-11 | evidence |
| GPT-5.3 Codex (xhigh) | OpenAI | 0.753741496598639 | 2026-02-05 | evidence |
| GPT-5.4 (low) | OpenAI | 0.659183673469388 | 2026-03-05 | evidence |
| GPT-5.4 (Non-reasoning) | OpenAI | 0.484353741496599 | 2026-03-05 | evidence |
| GPT-5.4 (xhigh) | OpenAI | 0.739455782312925 | 2026-03-05 | evidence |
| GPT-5.4 mini (medium) | OpenAI | 0.647619047619048 | 2026-03-17 | evidence |
| GPT-5.4 mini (Non-Reasoning) | OpenAI | 0.38843537414966 | 2026-03-17 | evidence |
| GPT-5.4 mini (xhigh) | OpenAI | 0.73265306122449 | 2026-03-17 | evidence |
| GPT-5.4 nano (medium) | OpenAI | 0.64421768707483 | 2026-03-17 | evidence |
| GPT-5.4 nano (Non-Reasoning) | OpenAI | 0.327210884353741 | 2026-03-17 | evidence |
| GPT-5.4 nano (xhigh) | OpenAI | 0.759183673469388 | 2026-03-17 | evidence |
| GPT-5.5 (high) | OpenAI | 0.716326530612245 | 2026-04-23 | evidence |
| GPT-5.5 (low) | OpenAI | 0.643537414965986 | 2026-04-23 | evidence |
| GPT-5.5 (medium) | OpenAI | 0.70952380952381 | 2026-04-23 | evidence |
| GPT-5.5 (Non-reasoning) | OpenAI | 0.461224489795918 | 2026-04-23 | evidence |
| GPT-5.5 (xhigh) | OpenAI | 0.758503401360544 | 2026-04-23 | evidence |
| GPT-5.5 Instant (May 2026) | OpenAI | 0.714965986394558 | 2026-05-05 | evidence |
| GPT-5.6 Sol (high) | OpenAI | 0.691836734693878 | 2026-07-09 | evidence |
| GPT-5.6 Sol (low) | OpenAI | 0.66530612244898 | 2026-07-09 | evidence |
| GPT-5.6 Sol (max) | OpenAI | 0.726530612244898 | 2026-07-09 | evidence |
| GPT-5.6 Sol (medium) | OpenAI | 0.695918367346939 | 2026-07-09 | evidence |
| GPT-5.6 Sol (xhigh) | OpenAI | 0.710204081632653 | 2026-07-09 | evidence |
| GPT-5.6 Terra (high) | OpenAI | 0.64421768707483 | 2026-07-09 | evidence |
| GPT-5.6 Terra (low) | OpenAI | 0.596598639455782 | 2026-07-09 | evidence |
| GPT-5.6 Terra (max) | OpenAI | 0.712244897959184 | 2026-07-09 | evidence |
| GPT-5.6 Terra (medium) | OpenAI | 0.621768707482993 | 2026-07-09 | evidence |
| GPT-5.6 Terra (xhigh) | OpenAI | 0.662585034013605 | 2026-07-09 | evidence |
| gpt-oss-120b (high) | OpenAI | 0.689795918367347 | 2025-08-05 | evidence |
| gpt-oss-120b (low) | OpenAI | 0.582993197278912 | 2025-08-05 | evidence |
| gpt-oss-20b (high) | OpenAI | 0.651020408163265 | 2025-08-05 | evidence |
| gpt-oss-20b (low) | OpenAI | 0.578231292517007 | 2025-08-05 | evidence |
| Granite 3.3 8B (Non-reasoning) | IBM | 0.224489795918367 | 2025-04-16 | evidence |
| Granite 4.0 1B | IBM | 0.205442176870748 | 2025-10-28 | evidence |
| Granite 4.0 350M | IBM | 0.159183673469388 | 2025-10-28 | evidence |
| Granite 4.0 H 1B | IBM | 0.261904761904762 | 2025-10-28 | evidence |
| Granite 4.0 H 350M | IBM | 0.175510204081633 | 2025-10-28 | evidence |
| Granite 4.0 H Small | IBM | 0.314965986394558 | 2025-09-22 | evidence |
| Granite 4.0 Micro | IBM | 0.247619047619048 | 2025-09-22 | evidence |
| Granite 4.1 30B | IBM | 0.443537414965986 | 2026-04-29 | evidence |
| Granite 4.1 3B | IBM | 0.336734693877551 | 2026-04-29 | evidence |
| Granite 4.1 8B | IBM | 0.386394557823129 | 2026-04-29 | evidence |
| Grok 3 | xAI | 0.469387755102041 | 2025-02-19 | evidence |
| Grok 3 mini Reasoning (high) | xAI | 0.458503401360544 | 2025-02-19 | evidence |
| Grok 4 | xAI | 0.536734693877551 | 2025-07-10 | evidence |
| Grok 4 Fast (Non-reasoning) | xAI | 0.37687074829932 | 2025-09-19 | evidence |
| Grok 4 Fast (Reasoning) | xAI | 0.505442176870748 | 2025-09-19 | evidence |
| Grok 4.1 Fast (Non-reasoning) | xAI | 0.364625850340136 | 2025-11-19 | evidence |
| Grok 4.1 Fast (Reasoning) | xAI | 0.527210884353742 | 2025-11-19 | evidence |
| Grok 4.20 0309 (Non-reasoning) | xAI | 0.478231292517007 | 2026-03-10 | evidence |
| Grok 4.20 0309 (Reasoning) | xAI | 0.829251700680272 | 2026-03-10 | evidence |
| Grok 4.20 0309 v2 (Non-reasoning) | xAI | 0.493197278911565 | 2026-04-07 | evidence |
| Grok 4.20 0309 v2 (Reasoning) | xAI | 0.812244897959184 | 2026-04-07 | evidence |
| Grok 4.3 (high) | xAI | 0.812925170068027 | 2026-04-30 | evidence |
| Grok 4.3 (low) | xAI | 0.80952380952381 | 2026-04-30 | evidence |
| Grok 4.3 (medium) | xAI | 0.833333333333333 | 2026-04-30 | evidence |
| Grok 4.3 (Non-reasoning) | xAI | 0.476190476190476 | 2026-04-30 | evidence |
| Grok Code Fast 1 | xAI | 0.413605442176871 | 2025-08-28 | evidence |
| Hermes 4 - Llama-3.1 405B (Non-reasoning) | Nous Research | 0.347619047619048 | 2025-08-27 | evidence |
| Hermes 4 - Llama-3.1 405B (Reasoning) | Nous Research | 0.327210884353742 | 2025-08-27 | evidence |
| Hermes 4 - Llama-3.1 70B (Non-reasoning) | Nous Research | 0.289795918367347 | 2025-08-27 | evidence |
| Hermes 4 - Llama-3.1 70B (Reasoning) | Nous Research | 0.312925170068027 | 2025-08-27 | evidence |
| Hy3-preview (Non-reasoning) | Tencent | 0.479591836734694 | 2026-04-23 | evidence |
| Hy3-preview (Reasoning) | Tencent | 0.631292517006803 | 2026-04-23 | evidence |
| HyperCLOVA X SEED Think (32B) | Naver | 0.37891156462585 | 2025-12-26 | evidence |
| HyperNova 60B 2605 | Multiverse Computing | 0.664625850340136 | 2026-05-26 | evidence |
| INTELLECT-3 | Prime Intellect | 0.340136054421769 | 2025-11-27 | evidence |
| Jamba 1.7 Large | AI21 Labs | 0.351700680272109 | 2025-07-07 | evidence |
| Jamba 1.7 Mini | AI21 Labs | 0.313605442176871 | 2025-07-07 | evidence |
| Jamba Reasoning 3B | AI21 Labs | 0.524489795918367 | 2025-10-08 | evidence |
| JT-35B-Flash | China Mobile | 0.420408163265306 | 2026-05-14 | evidence |
| JT-MINI | China Mobile | 0.366666666666667 | 2026-04-15 | evidence |
| K-EXAONE (Non-reasoning) | LG AI Research | 0.395918367346939 | 2025-12-31 | evidence |
| K-EXAONE (Reasoning) | LG AI Research | 0.646938775510204 | 2025-12-31 | evidence |
| K2 Think V2 | MBZUAI Institute of Foundation Models | 0.627891156462585 | 2025-12-15 | evidence |
| K2-V2 (high) | MBZUAI Institute of Foundation Models | 0.601360544217687 | 2025-12-05 | evidence |
| K2-V2 (low) | MBZUAI Institute of Foundation Models | 0.40952380952381 | 2025-12-05 | evidence |
| K2-V2 (medium) | MBZUAI Institute of Foundation Models | 0.551020408163265 | 2025-12-05 | evidence |
| KAT Coder Pro V2 | KwaiKAT | 0.666666666666667 | 2026-03-27 | evidence |
| KAT-Coder-Pro V1 | KwaiKAT | 0.683673469387755 | 2025-11-11 | evidence |
| Kimi K2 | Moonshot AI | 0.414965986394558 | 2025-07-11 | evidence |
| Kimi K2 0905 | Moonshot AI | 0.417006802721088 | 2025-09-05 | evidence |
| Kimi K2 Thinking | Moonshot AI | 0.680952380952381 | 2025-11-06 | evidence |
| Kimi K2.5 (Non-reasoning) | Moonshot AI | 0.436734693877551 | 2026-01-27 | evidence |
| Kimi K2.5 (Reasoning) | Moonshot AI | 0.702040816326531 | 2026-01-27 | evidence |
| Kimi K2.6 | Moonshot AI | 0.759863945578231 | 2026-04-20 | evidence |
| Kimi K2.6 (Non-reasoning) | Moonshot AI | 0.442857142857143 | 2026-04-20 | evidence |
| Kimi K2.7 Code | Moonshot AI | 0.631292517006803 | 2026-06-12 | evidence |
| Kimi Linear 48B A3B Instruct | Moonshot AI | 0.280952380952381 | 2025-10-30 | evidence |
| LFM2 1.2B | Liquid AI | 0.219727891156463 | 2025-07-10 | evidence |
| LFM2 2.6B | Liquid AI | 0.195238095238095 | 2025-09-23 | evidence |
| LFM2 24B A2B | Liquid AI | 0.458503401360544 | 2026-02-25 | evidence |
| LFM2 8B A1B | Liquid AI | 0.262585034013605 | 2025-10-07 | evidence |
| LFM2.5-1.2B-Instruct | Liquid AI | 0.438095238095238 | 2026-01-05 | evidence |
| LFM2.5-1.2B-Thinking | Liquid AI | 0.418367346938776 | 2026-01-20 | evidence |
| LFM2.5-8B-A1B | Liquid AI | 0.556462585034014 | 2026-05-28 | evidence |
| LFM2.5-VL-1.6B | Liquid AI | 0.331292517006803 | 2026-01-05 | evidence |
| Ling 2.6 Flash | InclusionAI | 0.574149659863946 | 2026-04-21 | evidence |
| Ling-1T | InclusionAI | 0.347619047619048 | 2025-10-08 | evidence |
| Ling-2.6-1T | InclusionAI | 0.568707482993197 | 2026-04-23 | evidence |
| Ling-flash-2.0 | InclusionAI | 0.343537414965986 | 2025-09-17 | evidence |
| Ling-mini-2.0 | InclusionAI | 0.236054421768707 | 2025-09-09 | evidence |
| Llama 3 Instruct 70B | Meta | 0.370748299319728 | 2024-04-18 | evidence |
| Llama 3 Instruct 8B | Meta | 0.245578231292517 | 2024-04-18 | evidence |
| Llama 3.1 Instruct 405B | Meta | 0.390476190476191 | 2024-07-23 | evidence |
| Llama 3.1 Instruct 70B | Meta | 0.34421768707483 | 2024-07-23 | evidence |
| Llama 3.1 Instruct 8B | Meta | 0.285714285714286 | 2024-07-23 | evidence |
| Llama 3.1 Nemotron Instruct 70B | NVIDIA | 0.307482993197279 | 2024-10-15 | evidence |
| Llama 3.1 Nemotron Nano 4B v1.1 (Reasoning) | NVIDIA | 0.255102040816327 | 2025-05-20 | evidence |
| Llama 3.1 Nemotron Ultra 253B v1 (Reasoning) | NVIDIA | 0.381632653061225 | 2025-04-07 | evidence |
| Llama 3.2 Instruct 11B (Vision) | Meta | 0.304081632653061 | 2024-09-25 | evidence |
| Llama 3.2 Instruct 1B | Meta | 0.227891156462585 | 2024-09-25 | evidence |
| Llama 3.2 Instruct 3B | Meta | 0.261904761904762 | 2024-09-25 | evidence |
| Llama 3.3 Instruct 70B | Meta | 0.470748299319728 | 2024-12-06 | evidence |
| Llama 3.3 Nemotron Super 49B v1 (Non-reasoning) | NVIDIA | 0.394557823129252 | 2025-03-18 | evidence |
| Llama 3.3 Nemotron Super 49B v1 (Reasoning) | NVIDIA | 0.380952380952381 | 2025-03-18 | evidence |
| Llama 4 Maverick | Meta | 0.429931972789116 | 2025-04-05 | evidence |
| Llama 4 Scout | Meta | 0.395238095238095 | 2025-04-05 | evidence |
| Llama Nemotron Super 49B v1.5 (Non-reasoning) | NVIDIA | 0.329251700680272 | 2025-07-25 | evidence |
| Llama Nemotron Super 49B v1.5 (Reasoning) | NVIDIA | 0.370068027210884 | 2025-07-25 | evidence |
| LongCat Flash Lite | LongCat | 0.430612244897959 | 2026-01-28 | evidence |
| Magistral Medium 1 | Mistral | 0.251020408163265 | 2025-06-10 | evidence |
| Magistral Medium 1.2 | Mistral | 0.429931972789116 | 2025-09-18 | evidence |
| Magistral Small 1 | Mistral | 0.247619047619048 | 2025-06-10 | evidence |
| Magistral Small 1.2 | Mistral | 0.443537414965986 | 2025-09-17 | evidence |
| Mercury 2 | Inception | 0.697959183673469 | 2026-02-20 | evidence |
| Mi:dm K 2.5 Pro | Korea Telecom | 0.493197278911565 | 2025-12-11 | evidence |
| Mi:dm K 2.5 Pro Preview | Korea Telecom | 0.45578231292517 | 2025-12-11 | evidence |
| MiMo-V2-Flash (Feb 2026) | Xiaomi | 0.718367346938776 | 2025-12-16 | evidence |
| MiMo-V2-Flash (Non-reasoning) | Xiaomi | 0.399319727891157 | 2025-12-16 | evidence |
| MiMo-V2-Flash (Reasoning) | Xiaomi | 0.642176870748299 | 2025-12-16 | evidence |
| MiMo-V2-Omni | Xiaomi | 0.535374149659864 | 2026-03-19 | evidence |
| MiMo-V2-Omni-0327 | Xiaomi | 0.673469387755102 | 2026-03-27 | evidence |
| MiMo-V2-Pro | Xiaomi | 0.68843537414966 | 2026-03-18 | evidence |
| MiMo-V2.5 | Xiaomi | 0.671428571428571 | 2026-04-22 | evidence |
| MiMo-V2.5-Pro | Xiaomi | 0.798639455782313 | 2026-04-22 | evidence |
| MiMo-V2.5-Pro (Non-reasoning) | Xiaomi | 0.427210884353741 | 2026-04-22 | evidence |
| MiniCPM-V 4.6 1.3B | OpenBMB | 0.26734693877551 | 2026-05-11 | evidence |
| MiniCPM5-1B (Non-reasoning) | OpenBMB | 0.351700680272109 | 2026-05-25 | evidence |
| MiniCPM5-1B (Reasoning) | OpenBMB | 0.493197278911565 | 2026-05-25 | evidence |
| MiniMax M1 40k | MiniMax | 0.412244897959184 | 2025-06-17 | evidence |
| MiniMax M1 80k | MiniMax | 0.417687074829932 | 2025-06-17 | evidence |
| MiniMax-M2 | MiniMax | 0.72312925170068 | 2025-10-26 | evidence |
| MiniMax-M2.1 | MiniMax | 0.698639455782313 | 2025-12-23 | evidence |
| MiniMax-M2.5 | MiniMax | 0.716326530612245 | 2026-02-12 | evidence |
| MiniMax-M2.7 | MiniMax | 0.757142857142857 | 2026-03-18 | evidence |
| MiniMax-M3 | MiniMax | 0.828571428571429 | 2026-06-01 | evidence |
| Ministral 3 14B | Mistral | 0.320408163265306 | 2025-12-02 | evidence |
| Ministral 3 3B | Mistral | 0.268027210884354 | 2025-12-02 | evidence |
| Ministral 3 8B | Mistral | 0.291156462585034 | 2025-12-02 | evidence |
| Mistral 7B Instruct | Mistral | 0.199319727891156 | 2023-09-27 | evidence |
| Mistral Large 2 (Jul '24) | Mistral | 0.316326530612245 | 2024-07-24 | evidence |
| Mistral Large 2 (Nov '24) | Mistral | 0.312244897959184 | 2024-11-18 | evidence |
| Mistral Large 3 | Mistral | 0.361904761904762 | 2025-12-02 | evidence |
| Mistral Medium 3 | Mistral | 0.392517006802721 | 2025-05-07 | evidence |
| Mistral Medium 3.1 | Mistral | 0.397959183673469 | 2025-08-12 | evidence |
| Mistral Medium 3.5 | Mistral | 0.687755102040816 | 2026-04-29 | evidence |
| Mistral Small 3 | Mistral | 0.263945578231293 | 2025-01-30 | evidence |
| Mistral Small 3.1 | Mistral | 0.299319727891156 | 2025-03-17 | evidence |
| Mistral Small 3.2 | Mistral | 0.33469387755102 | 2025-06-20 | evidence |
| Mistral Small 4 (Non-reasoning) | Mistral | 0.327891156462585 | 2026-03-16 | evidence |
| Mistral Small 4 (Reasoning) | Mistral | 0.481632653061224 | 2026-03-16 | evidence |
| Molmo 7B-D | Allen Institute for AI | 0.196598639455782 | 2024-09-25 | evidence |
| Molmo2-8B | Allen Institute for AI | 0.269387755102041 | 2025-12-11 | evidence |
| Motif-2-12.7B-Reasoning | Motif Technologies | 0.570068027210884 | 2025-12-04 | evidence |
| Muse Spark | Meta | 0.759183673469388 | 2026-04-08 | evidence |
| Nanbeige4.1-3B | Nanbeige | 0.354421768707483 | 2026-02-11 | evidence |
| Nemotron 3 Nano Omni 30B A3B Reasoning | NVIDIA | 0.631972789115646 | 2026-04-29 | evidence |
| Nemotron 3 Super 120B A12B (Reasoning) | NVIDIA | 0.714965986394558 | 2026-03-11 | evidence |
| Nemotron 3 Ultra 550B A55B (Reasoning) | NVIDIA | 0.813605442176871 | 2026-06-04 | evidence |
| Nemotron Cascade 2 30B A3B | NVIDIA | 0.804081632653061 | 2026-03-19 | evidence |
| Nex-N2-Pro | Nex AGI | 0.661904761904762 | 2026-06-02 | evidence |
| North Mini Code | Cohere | 0.575510204081633 | 2026-06-09 | evidence |
| Nova 2.0 Lite (high) | Amazon | 0.707482993197279 | 2025-10-29 | evidence |
| Nova 2.0 Lite (low) | Amazon | 0.612244897959184 | 2025-10-29 | evidence |
| Nova 2.0 Lite (medium) | Amazon | 0.685034013605442 | 2025-10-29 | evidence |
| Nova 2.0 Lite (Non-reasoning) | Amazon | 0.405442176870748 | 2025-10-29 | evidence |
| Nova 2.0 Omni (low) | Amazon | 0.617687074829932 | 2025-11-26 | evidence |
| Nova 2.0 Omni (medium) | Amazon | 0.661904761904762 | 2025-11-26 | evidence |
| Nova 2.0 Omni (Non-reasoning) | Amazon | 0.410884353741497 | 2025-11-26 | evidence |
| Nova 2.0 Pro Preview (low) | Amazon | 0.795918367346939 | 2025-11-27 | evidence |
| Nova 2.0 Pro Preview (medium) | Amazon | 0.790476190476191 | 2025-11-27 | evidence |
| Nova 2.0 Pro Preview (Non-reasoning) | Amazon | 0.520408163265306 | 2025-11-27 | evidence |
| Nova Lite | Amazon | 0.341496598639456 | 2024-12-03 | evidence |
| Nova Micro | Amazon | 0.293877551020408 | 2024-12-03 | evidence |
| Nova Premier | Amazon | 0.361904761904762 | 2025-04-30 | evidence |
| Nova Pro | Amazon | 0.380952380952381 | 2024-12-03 | evidence |
| NVIDIA Nemotron 3 Nano 30B A3B (Non-reasoning) | NVIDIA | 0.374829931972789 | 2025-12-15 | evidence |
| NVIDIA Nemotron 3 Nano 30B A3B (Reasoning) | NVIDIA | 0.710884353741497 | 2025-12-15 | evidence |
| NVIDIA Nemotron 3 Nano 4B | NVIDIA | 0.582312925170068 | 2026-03-16 | evidence |
| NVIDIA Nemotron Nano 12B v2 VL (Non-reasoning) | NVIDIA | 0.258503401360544 | 2025-10-28 | evidence |
| NVIDIA Nemotron Nano 12B v2 VL (Reasoning) | NVIDIA | 0.319047619047619 | 2025-10-28 | evidence |
| NVIDIA Nemotron Nano 9B V2 (Non-reasoning) | NVIDIA | 0.271428571428571 | 2025-08-18 | evidence |
| NVIDIA Nemotron Nano 9B V2 (Reasoning) | NVIDIA | 0.276190476190476 | 2025-08-18 | evidence |
| o1 | OpenAI | 0.703401360544218 | 2024-12-05 | evidence |
| o3 | OpenAI | 0.714285714285714 | 2025-04-16 | evidence |
| o3-mini (high) | OpenAI | 0.671428571428571 | 2025-01-31 | evidence |
| o4-mini (high) | OpenAI | 0.687074829931973 | 2025-04-16 | evidence |
| OLMo 2 32B | Allen Institute for AI | 0.380952380952381 | 2025-03-13 | evidence |
| OLMo 2 7B | Allen Institute for AI | 0.24421768707483 | 2024-11-26 | evidence |
| Olmo 3 32B Think | Allen Institute for AI | 0.491156462585034 | 2025-11-20 | evidence |
| Olmo 3 7B Instruct | Allen Institute for AI | 0.327891156462585 | 2025-11-20 | evidence |
| Olmo 3 7B Think | Allen Institute for AI | 0.414965986394558 | 2025-11-20 | evidence |
| Olmo 3.1 32B Instruct | Allen Institute for AI | 0.391836734693878 | 2026-01-13 | evidence |
| Olmo 3.1 32B Think | Allen Institute for AI | 0.659863945578231 | 2025-12-12 | evidence |
| Phi-3 Mini Instruct 3.8B | Microsoft | 0.238775510204082 | 2024-04-23 | evidence |
| Phi-4 | Microsoft | 0.235374149659864 | 2024-12-12 | evidence |
| Phi-4 Mini Instruct | Microsoft | 0.210884353741497 | 2024-02-26 | evidence |
| Pixtral Large | Mistral | 0.344897959183673 | 2024-11-18 | evidence |
| Qwen2.5 Instruct 72B | Qwen | 0.368707482993197 | 2024-09-19 | evidence |
| Qwen3 0.6B (Non-reasoning) | Qwen | 0.219047619047619 | 2025-04-28 | evidence |
| Qwen3 0.6B (Reasoning) | Qwen | 0.233333333333333 | 2025-04-28 | evidence |
| Qwen3 1.7B (Non-reasoning) | Qwen | 0.210884353741497 | 2025-04-28 | evidence |
| Qwen3 1.7B (Reasoning) | Qwen | 0.268707482993197 | 2025-04-28 | evidence |
| Qwen3 14B (Non-reasoning) | Qwen | 0.239455782312925 | 2025-04-28 | evidence |
| Qwen3 14B (Reasoning) | Qwen | 0.405442176870748 | 2025-04-28 | evidence |
| Qwen3 235B A22B (Non-reasoning) | Qwen | 0.365986394557823 | 2025-04-28 | evidence |
| Qwen3 235B A22B (Reasoning) | Qwen | 0.387074829931973 | 2025-04-28 | evidence |
| Qwen3 235B A22B 2507 (Reasoning) | Qwen | 0.512244897959184 | 2025-07-25 | evidence |
| Qwen3 235B A22B 2507 Instruct | Qwen | 0.460544217687075 | 2025-07-21 | evidence |
| Qwen3 30B A3B (Non-reasoning) | Qwen | 0.319047619047619 | 2025-04-28 | evidence |
| Qwen3 30B A3B (Reasoning) | Qwen | 0.414965986394558 | 2025-04-28 | evidence |
| Qwen3 30B A3B 2507 (Reasoning) | Qwen | 0.506802721088435 | 2025-07-30 | evidence |
| Qwen3 30B A3B 2507 Instruct | Qwen | 0.330612244897959 | 2025-07-29 | evidence |
| Qwen3 32B (Non-reasoning) | Qwen | 0.314965986394558 | 2025-04-28 | evidence |
| Qwen3 32B (Reasoning) | Qwen | 0.363265306122449 | 2025-04-28 | evidence |
| Qwen3 4B (Reasoning) | Qwen | 0.325170068027211 | 2025-04-28 | evidence |
| Qwen3 4B 2507 (Reasoning) | Qwen | 0.497959183673469 | 2025-08-06 | evidence |
| Qwen3 4B 2507 Instruct | Qwen | 0.335374149659864 | 2025-08-06 | evidence |
| Qwen3 8B (Non-reasoning) | Qwen | 0.285714285714286 | 2025-04-28 | evidence |
| Qwen3 8B (Reasoning) | Qwen | 0.33469387755102 | 2025-04-28 | evidence |
| Qwen3 Coder 30B A3B Instruct | Qwen | 0.326530612244898 | 2025-07-31 | evidence |
| Qwen3 Coder 480B A35B Instruct | Qwen | 0.404761904761905 | 2025-07-22 | evidence |
| Qwen3 Coder Next | Qwen | 0.352380952380952 | 2026-02-03 | evidence |
| Qwen3 Max | Qwen | 0.441496598639456 | 2025-09-23 | evidence |
| Qwen3 Max (Preview) | Qwen | 0.480272108843537 | 2025-09-05 | evidence |
| Qwen3 Max Thinking | Qwen | 0.707482993197279 | 2026-01-26 | evidence |
| Qwen3 Max Thinking (Preview) | Qwen | 0.538095238095238 | 2025-11-03 | evidence |
| Qwen3 Next 80B A3B (Reasoning) | Qwen | 0.606802721088435 | 2025-09-11 | evidence |
| Qwen3 Next 80B A3B Instruct | Qwen | 0.396598639455782 | 2025-09-11 | evidence |
| Qwen3 Omni 30B A3B (Reasoning) | Qwen | 0.434013605442177 | 2025-09-22 | evidence |
| Qwen3 Omni 30B A3B Instruct | Qwen | 0.31156462585034 | 2025-09-22 | evidence |
| Qwen3 VL 235B A22B (Reasoning) | Qwen | 0.564625850340136 | 2025-09-23 | evidence |
| Qwen3 VL 235B A22B Instruct | Qwen | 0.426530612244898 | 2025-09-23 | evidence |
| Qwen3 VL 30B A3B (Reasoning) | Qwen | 0.451020408163265 | 2025-10-03 | evidence |
| Qwen3 VL 30B A3B Instruct | Qwen | 0.331292517006803 | 2025-10-03 | evidence |
| Qwen3 VL 32B (Reasoning) | Qwen | 0.593877551020408 | 2025-10-21 | evidence |
| Qwen3 VL 32B Instruct | Qwen | 0.391836734693878 | 2025-10-21 | evidence |
| Qwen3 VL 4B (Reasoning) | Qwen | 0.365986394557823 | 2025-10-14 | evidence |
| Qwen3 VL 4B Instruct | Qwen | 0.318367346938775 | 2025-10-14 | evidence |
| Qwen3 VL 8B (Reasoning) | Qwen | 0.398639455782313 | 2025-10-14 | evidence |
| Qwen3 VL 8B Instruct | Qwen | 0.32312925170068 | 2025-10-14 | evidence |
| Qwen3.5 0.8B (Non-reasoning) | Qwen | 0.215646258503401 | 2026-03-02 | evidence |
| Qwen3.5 0.8B (Reasoning) | Qwen | 0.214965986394558 | 2026-03-02 | evidence |
| Qwen3.5 122B A10B (Non-reasoning) | Qwen | 0.508163265306122 | 2026-02-24 | evidence |
| Qwen3.5 122B A10B (Reasoning) | Qwen | 0.757142857142857 | 2026-02-24 | evidence |
| Qwen3.5 27B (Non-reasoning) | Qwen | 0.469387755102041 | 2026-02-24 | evidence |
| Qwen3.5 27B (Reasoning) | Qwen | 0.75578231292517 | 2026-02-24 | evidence |
| Qwen3.5 2B (Non-reasoning) | Qwen | 0.291156462585034 | 2026-03-02 | evidence |
| Qwen3.5 2B (Reasoning) | Qwen | 0.314965986394558 | 2026-03-02 | evidence |
| Qwen3.5 35B A3B (Non-reasoning) | Qwen | 0.444897959183673 | 2026-02-24 | evidence |
| Qwen3.5 35B A3B (Reasoning) | Qwen | 0.725170068027211 | 2026-02-24 | evidence |
| Qwen3.5 397B A17B (Non-reasoning) | Qwen | 0.516326530612245 | 2026-02-16 | evidence |
| Qwen3.5 397B A17B (Reasoning) | Qwen | 0.787755102040816 | 2026-02-16 | evidence |
| Qwen3.5 4B (Non-reasoning) | Qwen | 0.33265306122449 | 2026-03-02 | evidence |
| Qwen3.5 4B (Reasoning) | Qwen | 0.519727891156463 | 2026-03-02 | evidence |
| Qwen3.5 9B (Non-reasoning) | Qwen | 0.378231292517007 | 2026-03-02 | evidence |
| Qwen3.5 9B (Reasoning) | Qwen | 0.66734693877551 | 2026-03-02 | evidence |
| Qwen3.5 Omni Flash | Qwen | 0.379591836734694 | 2026-03-30 | evidence |
| Qwen3.5 Omni Plus | Qwen | 0.51156462585034 | 2026-03-30 | evidence |
| Qwen3.6 27B (Non-reasoning) | Qwen | 0.457142857142857 | 2026-04-22 | evidence |
| Qwen3.6 27B (Reasoning) | Qwen | 0.675510204081633 | 2026-04-22 | evidence |
| Qwen3.6 35B A3B (Non-reasoning) | Qwen | 0.361904761904762 | 2026-04-16 | evidence |
| Qwen3.6 35B A3B (Reasoning) | Qwen | 0.643537414965986 | 2026-04-16 | evidence |
| Qwen3.6 Max Preview | Qwen | 0.765986394557823 | 2026-04-20 | evidence |
| Qwen3.6 Plus | Qwen | 0.751700680272109 | 2026-04-02 | evidence |
| Qwen3.7 Max | Qwen | 0.805442176870748 | 2026-05-19 | evidence |
| Qwen3.7 Plus | Qwen | 0.779591836734694 | 2026-06-01 | evidence |
| QwQ 32B | Qwen | 0.387755102040816 | 2025-03-05 | evidence |
| Reka Flash 3 | Reka AI | 0.304081632653061 | 2025-03-10 | evidence |
| Ring-1T | InclusionAI | 0.445578231292517 | 2025-10-13 | evidence |
| Ring-2.6-1T | InclusionAI | 0.445578231292517 | 2026-05-08 | evidence |
| Ring-flash-2.0 | InclusionAI | 0.43265306122449 | 2025-09-19 | evidence |
| Sarvam 105B (high) | Sarvam | 0.343537414965986 | 2026-03-06 | evidence |
| Sarvam 30B (high) | Sarvam | 0.264625850340136 | 2026-03-06 | evidence |
| Sarvam M (Reasoning) | Sarvam | 0.318367346938775 | 2025-05-23 | evidence |
| Seed-OSS-36B-Instruct | ByteDance Seed | 0.419047619047619 | 2025-08-20 | evidence |
| Solar Open 100B (Reasoning) | Upstage | 0.57687074829932 | 2025-12-17 | evidence |
| Solar Pro 2 (Non-reasoning) | Upstage | 0.337414965986395 | 2025-07-09 | evidence |
| Solar Pro 2 (Reasoning) | Upstage | 0.370748299319728 | 2025-07-09 | evidence |
| Solar Pro 3 | Upstage | 0.712018140589569 | 2026-04-06 | evidence |
| Step 3.5 Flash | StepFun | 0.645578231292517 | 2026-02-02 | evidence |
| Step 3.5 Flash 2603 | StepFun | 0.66530612244898 | 2026-04-02 | evidence |
| Step 3.7 Flash | StepFun | 0.672789115646258 | 2026-05-29 | evidence |
| Step3 VL 10B | StepFun | 0.502040816326531 | 2026-01-20 | evidence |
| Tiny Aya Global | Cohere | 0.201360544217687 | 2026-02-17 | evidence |
| Tri-21B-Think | Trillion Labs | 0.546258503401361 | 2026-02-10 | evidence |
| Tri-21B-think Preview | Trillion Labs | 0.471428571428571 | 2026-02-10 | evidence |
| Trinity Large Thinking | Arcee AI | 0.562585034013605 | 2026-04-01 | evidence |
Scores are partitioned by the source that reported them and are never merged into a single cross-source ranking, because the sources measure different things and say so.