Skip to content
ModelCap

Charts

AI model price vs performance, and the frontier over time

Two views of the ModelCap Index, drawn from the same numbers as the main board. The first sets every current ranked model that lists an API price against its Index score and traces the best-value line: the highest measured score available at or below each price. The second follows the best measured score by the date models were first listed.

Snapshot as of 23 September 2026

Capability vs price

115 of 274 current ranked models list an API price. The best-value line runs through 12 measured models, from gpt-oss-20b ($0.036 per million tokens, score 39.1) to Claude Opus 5.5 ($8.00, score 97.1).

ModelCap Index score against blended API price113 current ranked models. The best-value line joins 12 measured models, each the highest measured score at or below its price.0102030405060708090100$0.01$0.10$1$10$100ModelCap Index scoreBlended API price, US$ per million tokens (log scale)Ling 3.0 Flash · #78 · score 60.0 (estimated) · $0.032 per 1M tokens blendedNex-N2.5-Mini · #31 · score 73.7 (estimated) · $0.044 per 1M tokens blendedLlama 3.1 8B Instruct · #227 · score 14.3 (measured) · $0.057 per 1M tokens blendedNova Micro 1.0 · #218 · score 18.4 (measured) · $0.061 per 1M tokens blendedGemma 3 4B · #185 · score 28.8 (measured) · $0.063 per 1M tokens blendedLlama 3.2 1B Instruct · #269 · score 5.4 (measured) · $0.071 per 1M tokens blendedLaguna XS 2.1 · #82 · score 58.6 (estimated) · $0.075 per 1M tokens blendedGemma 3 12B · #160 · score 39.7 (measured) · $0.075 per 1M tokens blendedQwen3 30B A3B Instruct 2507 · #109 · score 52.7 (measured) · $0.084 per 1M tokens blendedNemotron 3 Nano 30B A3B · #173 · score 33.4 (measured) · $0.087 per 1M tokens blendedPhi 4 · #213 · score 20.6 (measured) · $0.088 per 1M tokens blendedLing 3.0 Flash Fin · #58 · score 66.0 (measured) · $0.09 per 1M tokens blendedGranite 4.2 8B · #168 · score 35.8 (measured) · $0.107 per 1M tokens blendedNemotron 3.5 Lightning · #88 · score 57.3 (measured) · $0.11 per 1M tokens blendedQwen3.5-9B · #119 · score 49.1 (estimated) · $0.112 per 1M tokens blendedLaguna S 2.1 · #39 · score 70.7 (estimated) · $0.113 per 1M tokens blendedNex-N2.5-Pro · #23 · score 79.5 (estimated) · $0.119 per 1M tokens blendedLlama 3.2 3B Instruct · #256 · score 9.0 (measured) · $0.12 per 1M tokens blendedQwen3 Coder 30B A3B Instruct · #94 · score 56.5 (estimated) · $0.123 per 1M tokens blendedQwen3 32B · #138 · score 43.2 (measured) · $0.13 per 1M tokens blendedgpt-oss-safeguard-20b · #170 · score 35.1 (inherited) · $0.131 per 1M tokens blendedMistral Small 3.2 24B · #135 · score 45.1 (measured) · $0.133 per 1M tokens blendedLlama 4 Scout · #171 · score 34.5 (measured) · $0.15 per 1M tokens blendedQwen3 235B A22B Instruct 2507 · #60 · score 64.7 (measured) · $0.153 per 1M tokens blendedLlama 3.3 70B Instruct · #172 · score 33.9 (measured) · $0.155 per 1M tokens blendedSolar Pro 4 · #113 · score 50.7 (measured) · $0.158 per 1M tokens blendedNemotron 3 Super · #128 · score 46.6 (measured) · $0.172 per 1M tokens blendedGemma 3 27B · #129 · score 46.5 (measured) · $0.172 per 1M tokens blendedMiMo-V2.6-Flash · #10 · score 83.9 (estimated) · $0.175 per 1M tokens blendedMiMo-V2.5 · #43 · score 70.3 (measured) · $0.175 per 1M tokens blendedGPT-6 Luna · #27 · score 78.4 (measured) · $0.20 per 1M tokens blendedQwen3 30B A3B · #165 · score 37.7 (measured) · $0.215 per 1M tokens blendedMistral Small 4 · #99 · score 55.8 (measured) · $0.262 per 1M tokens blendedgpt-oss-120b · #130 · score 46.3 (measured) · $0.262 per 1M tokens blendedGPT-4o-mini · #156 · score 40.3 (measured) · $0.262 per 1M tokens blendedCommand R (08-2024) · #215 · score 19.5 (measured) · $0.262 per 1M tokens blendedQwen3 Coder Next · #107 · score 54.0 (measured) · $0.29 per 1M tokens blendedDeepSeek V3.2 · #61 · score 64.4 (measured) · $0.302 per 1M tokens blendedLlama 4 Maverick · #167 · score 36.3 (measured) · $0.304 per 1M tokens blendedGLM 4.5 Air · #114 · score 50.5 (measured) · $0.31 per 1M tokens blendedQwen3 Next 80B A3B Instruct · #89 · score 57.2 (measured) · $0.343 per 1M tokens blendedQwen3.6 35B A3B · #73 · score 62.1 (measured) · $0.362 per 1M tokens blendedQwen2.5 72B Instruct · #183 · score 30.7 (measured) · $0.37 per 1M tokens blendedMistral Large 3 2512 · #71 · score 62.2 (measured) · $0.375 per 1M tokens blendedTrinity Large Thinking · #115 · score 50.4 (measured) · $0.387 per 1M tokens blendedQwen3 Next 80B A3B Thinking · #117 · score 49.3 (measured) · $0.412 per 1M tokens blendedMiniMax-01 · #136 · score 43.9 (estimated) · $0.425 per 1M tokens blendedStep 3.7 Flash · #98 · score 56.0 (estimated) · $0.438 per 1M tokens blendedGLM 4.6V · #112 · score 50.9 (measured) · $0.45 per 1M tokens blendedDeepSeek V3.1 Terminus · #70 · score 62.4 (measured) · $0.453 per 1M tokens blendedGPT-5.4 Nano · #80 · score 58.8 (measured) · $0.463 per 1M tokens blendedQwen3 Coder 480B A35B · #105 · score 54.3 (measured) · $0.475 per 1M tokens blendedLongCat 2.0 · #16 · score 81.4 (estimated) · $0.525 per 1M tokens blendedMuse Glimmer 30B · #36 · score 71.5 (measured) · $0.525 per 1M tokens blendedMiniMax M3 · #37 · score 71.3 (measured) · $0.525 per 1M tokens blendedQwen3.7 Plus · #30 · score 73.7 (measured) · $0.56 per 1M tokens blendedCommand A+ · #87 · score 57.5 (measured) · $0.60 per 1M tokens blendedQwen3 VL 235B A22B Instruct · #64 · score 62.8 (measured) · $0.632 per 1M tokens blendedInkling Small · #84 · score 58.1 (measured) · $0.637 per 1M tokens blendedDeepSeek V4 Pro 0813 · #25 · score 78.5 (measured) · $0.693 per 1M tokens blendedQwen3.5-122B-A10B · #72 · score 62.1 (measured) · $0.715 per 1M tokens blendedQwen2.5 Coder 32B Instruct · #204 · score 24.8 (measured) · $0.745 per 1M tokens blendedQwen3 235B A22B Thinking 2507 · #90 · score 57.2 (measured) · $0.747 per 1M tokens blendedQwen3 235B A22B · #86 · score 57.7 (measured) · $0.796 per 1M tokens blendedGemini 3.5 Flash Lite · #33 · score 72.6 (measured) · $0.85 per 1M tokens blendedNova 2 Lite · #155 · score 40.4 (measured) · $0.85 per 1M tokens blendedR1 0528 · #62 · score 64.3 (measured) · $0.913 per 1M tokens blendedSeed 2.1 Turbo · #21 · score 79.6 (estimated) · $1.00 per 1M tokens blendedQwen3.8 27B · #35 · score 71.5 (measured) · $1.06 per 1M tokens blendedKimi K2 Thinking · #47 · score 68.5 (measured) · $1.07 per 1M tokens blendedHy4 preview · #28 · score 76.9 (measured) · $1.25 per 1M tokens blendedQwen3.5 397B A17B · #42 · score 70.4 (measured) · $1.29 per 1M tokens blendedGLM 5.3 · #11 · score 83.8 (measured) · $1.29 per 1M tokens blendedQwen3 VL 235B A22B Thinking · #95 · score 56.4 (measured) · $1.30 per 1M tokens blendedKimi K2.7 Code · #50 · score 67.7 (measured) · $1.35 per 1M tokens blendedNova Pro 1.0 · #189 · score 27.9 (measured) · $1.40 per 1M tokens blendedGemini 3.8 Flash · #9 · score 84.1 (measured) · $1.50 per 1M tokens blendedGrok 4.20 Multi-Agent · #18 · score 81.0 (measured) · $1.56 per 1M tokens blendedGPT-5.4 Mini · #34 · score 72.3 (measured) · $1.69 per 1M tokens blendedInkling · #46 · score 69.2 (measured) · $1.76 per 1M tokens blendedGLM 5V Turbo · #45 · score 69.6 (measured) · $1.90 per 1M tokens blendedGLM 5 Turbo · #103 · score 54.7 (measured) · $1.90 per 1M tokens blendedo4 Mini · #108 · score 53.6 (measured) · $1.93 per 1M tokens blendedMuse Spark 1.3 · #8 · score 85.6 (measured) · $2.00 per 1M tokens blendedClaude Haiku 4.5 · #65 · score 62.6 (measured) · $2.00 per 1M tokens blendedQwen3.8 Max (0902) · #6 · score 86.5 (measured) · $3.00 per 1M tokens blendedMistral Medium 3.5 · #63 · score 63.7 (measured) · $3.00 per 1M tokens blendedMixtral 8x22B Instruct · #223 · score 16.6 (measured) · $3.00 per 1M tokens blendedo3 · #52 · score 67.4 (measured) · $3.50 per 1M tokens blendedClaude Sonnet 5 · #26 · score 78.4 (measured) · $4.00 per 1M tokens blendedCommand A · #137 · score 43.8 (measured) · $4.38 per 1M tokens blendedGPT-4o (2024-08-06) · #162 · score 39.4 (measured) · $4.38 per 1M tokens blendedCommand R+ (08-2024) · #205 · score 24.3 (measured) · $4.38 per 1M tokens blendedGemini 3.1 Pro Preview · #17 · score 81.2 (measured) · $4.50 per 1M tokens blendedGPT-5.6 Terra · #22 · score 79.5 (measured) · $4.50 per 1M tokens blendedKimi K3 · #12 · score 83.5 (measured) · $6.00 per 1M tokens blendedGPT-5.5 · #13 · score 82.2 (measured) · $11.25 per 1M tokens blendedGPT-4 Turbo · #169 · score 35.2 (measured) · $15.00 per 1M tokens blendedClaude Fable 5.1 · #3 · score 87.8 (measured) · $20.00 per 1M tokens blendedGPT-6 Astra · #7 · score 86.2 (measured) · $20.00 per 1M tokens blendedGPT-5.5 Pro · #59 · score 65.3 (measured) · $67.50 per 1M tokens blendedgpt-oss-20b · #163 · score 39.1 (measured) · $0.036 per 1M tokens blendedMercury 2.5 · #93 · score 56.8 (measured) · $0.068 per 1M tokens blendedLing 3.0 Flash VL · #49 · score 67.8 (measured) · $0.09 per 1M tokens blendedGemma 4 26B A4B · #38 · score 70.8 (measured) · $0.143 per 1M tokens blendedGemma 4 31B · #29 · score 74.1 (measured) · $0.152 per 1M tokens blendedDeepSeek V4.1 Flash · #19 · score 80.2 (measured) · $0.20 per 1M tokens blendedQwen3.8 Flash · #15 · score 81.5 (measured) · $0.23 per 1M tokens blendedGLM 5.3 Flash · #14 · score 81.6 (measured) · $0.237 per 1M tokens blendedMiMo-V2.6-Pro · #5 · score 87.3 (measured) · $0.544 per 1M tokens blendedGrok 4.7 · #4 · score 87.4 (measured) · $2.40 per 1M tokens blendedGPT-6 Sol · #2 · score 88.4 (measured) · $4.00 per 1M tokens blendedClaude Opus 5.5 · #1 · score 97.1 (measured) · $8.00 per 1M tokens blendedClaude Opus 5.5GPT-6 SolGrok 4.7MiMo-V2.6-ProGLM 5.3 FlashDeepSeek V4.1 FlashGemma 4 31BGemma 4 26B A4BLing 3.0 Flash VLMercury 2.5gpt-oss-20b
Measured score Inherited or estimated score Best measured score at or below each price

Prices are the listed API prices from OpenRouter, blended three parts input to one part output, in US dollars per million tokens. The price axis is logarithmic: each gridline is ten times the one before. Inherited and estimated scores are modeled rather than measured, so they are drawn but never form the line. Models with no listed API price, including open-weight models you can only run yourself, are not drawn. Listed free of charge, and so off the logarithmic axis: Dots3-Note Preview (61.5) and North Mini Code (54.6).

The best-value line, cheapest first

Measured models on the best-value line, cheapest first
ModelIndex scorePositionBlended / 1MInput · output / 1M
gpt-oss-20bOpenAI39.1#163$0.036$0.018 · $0.09
Mercury 2.5Inception Labs56.8Preliminary#93$0.068$0.04 · $0.15
Ling 3.0 Flash VLInclusionAI67.8Preliminary#49$0.09$0.06 · $0.18
Gemma 4 26B A4BGoogle70.8#38$0.143$0.09 · $0.30
Gemma 4 31BGoogle74.1#29$0.152$0.09 · $0.34
DeepSeek V4.1 FlashDeepSeek80.2Preliminary#19$0.20$0.10 · $0.50
Qwen3.8 FlashQwen81.5Preliminary#15$0.23$0.15 · $0.47
GLM 5.3 FlashZ.ai81.6#14$0.237$0.15 · $0.50
MiMo-V2.6-ProXiaomi87.3Preliminary#5$0.544$0.435 · $0.87
Grok 4.7xAI87.4Preliminary#4$2.40$1.60 · $4.80
GPT-6 SolOpenAI88.4Preliminary#2$4.00$2.00 · $10.00
Claude Opus 5.5Anthropic97.1Preliminary#1$8.00$4.00 · $20.00

The frontier over time

17 models have each set a new best measured score when listed, from GPT-4 (27.2, listed 28 May 2023) to Claude Opus 5.5 (97.1, listed 22 Sept 2026).

Best measured ModelCap Index score by listing date235 measured models placed at the date each was first listed. The line steps up at each of the 17 models that set a new best score.0102030405060708090100202420252026ModelCap Index scoreDate first listedGPT-3.5 Turbo · listed 28 May 2023 · score 16.5 · all versions #340Qwen 14B Chat · listed 24 Sept 2023 · score 7.6 · all versions #396Mistral 7B Instruct v0.2 · listed 11 Dec 2023 · score 7.6 · #265Qwen1.5 4B Chat · listed 30 Jan 2024 · score 4.2 · #271Qwen1.5 7B Chat · listed 30 Jan 2024 · score 7.5 · #266Qwen1.5 14B Chat · listed 30 Jan 2024 · score 12.1 · #248Qwen1.5 72B Chat · listed 30 Jan 2024 · score 17.2 · #222OLMo 7B Instruct · listed 23 Feb 2024 · score 4.0 · #274Claude 3 Haiku · listed 13 Mar 2024 · score 21.8 · all versions #325gemma 1.1 7b it · listed 26 Mar 2024 · score 10.8 · #253gemma 1.1 2b it · listed 26 Mar 2024 · score 6.0 · all versions #402Qwen1.5 32B Chat · listed 3 Apr 2024 · score 13.5 · #237Mixtral 8x22B Instruct · listed 17 Apr 2024 · score 16.6 · #223Phi 3 mini 4k instruct · listed 22 Apr 2024 · score 7.0 · all versions #399Phi 3 mini 128k instruct · listed 22 Apr 2024 · score 6.8 · all versions #400Phi 3 medium 4k instruct · listed 7 May 2024 · score 12.7 · #243Phi 3 small 8k instruct · listed 7 May 2024 · score 10.0 · #254GPT-4o · listed 13 May 2024 · score 40.5 · all versions #251Qwen2 72B Instruct · listed 28 May 2024 · score 21.3 · all versions #327gemma 2 9b it · listed 24 Jun 2024 · score 21.8 · #212Gemma 2 27B · listed 13 Jul 2024 · score 26.9 · all versions #300gemma 2 2b it · listed 16 Jul 2024 · score 12.2 · #247GPT-4o-mini · listed 18 Jul 2024 · score 40.3 · #156GPT-4o-mini (2024-07-18) · listed 18 Jul 2024 · score 34.1 · all versions #272Llama 3.1 70B Instruct · listed 23 Jul 2024 · score 28.0 · all versions #292Llama 3.1 8B Instruct · listed 23 Jul 2024 · score 14.3 · #227GPT-4o (2024-08-06) · listed 6 Aug 2024 · score 39.4 · #162Command R+ (08-2024) · listed 30 Aug 2024 · score 24.3 · #205Command R (08-2024) · listed 30 Aug 2024 · score 19.5 · #215Qwen2.5 72B Instruct · listed 19 Sept 2024 · score 30.7 · #183Llama 3 1 Nemotron 51B Instruct · listed 22 Sept 2024 · score 26.3 · #195Llama 3.2 3B Instruct · listed 25 Sept 2024 · score 9.0 · #256Llama 3.2 1B Instruct · listed 25 Sept 2024 · score 5.4 · #269granite 3.0 2b instruct · listed 2 Oct 2024 · score 8.6 · all versions #383granite 3.0 8b instruct · listed 2 Oct 2024 · score 11.1 · all versions #373Qwen2.5 Coder 32B Instruct · listed 11 Nov 2024 · score 24.8 · #204Mistral Large 2407 · listed 19 Nov 2024 · score 33.5 · all versions #274Llama 3.1 Tulu 3 8B · listed 20 Nov 2024 · score 15.6 · #225Llama 3.1 Tulu 3 70B · listed 20 Nov 2024 · score 26.2 · #198QwQ 32B Preview · listed 27 Nov 2024 · score 8.3 · all versions #390Nova Pro 1.0 · listed 5 Dec 2024 · score 27.9 · #189Nova Micro 1.0 · listed 5 Dec 2024 · score 18.4 · #218Nova Lite 1.0 · listed 5 Dec 2024 · score 21.6 · all versions #326Llama 3.3 70B Instruct · listed 6 Dec 2024 · score 33.9 · #172granite 3.1 8b instruct · listed 6 Dec 2024 · score 14.4 · all versions #343granite 3.1 2b instruct · listed 6 Dec 2024 · score 11.1 · all versions #374DeepSeek V3 · listed 26 Dec 2024 · score 46.8 · all versions #212Phi 4 · listed 10 Jan 2025 · score 20.6 · #213R1 · listed 20 Jan 2025 · score 56.3 · all versions #164Mistral Small 3 · listed 30 Jan 2025 · score 24.8 · all versions #314o3 Mini · listed 31 Jan 2025 · score 42.8 · all versions #230Qwen-Plus · listed 1 Feb 2025 · score 42.4 · all versions #237QwQ 32B · listed 5 Mar 2025 · score 40.4 · #154Gemma 3 27B · listed 12 Mar 2025 · score 46.5 · #129OLMo 2 0325 32B Instruct · listed 12 Mar 2025 · score 19.8 · #214Command A · listed 13 Mar 2025 · score 43.8 · #137Gemma 3 12B · listed 13 Mar 2025 · score 39.7 · #160Gemma 3 4B · listed 13 Mar 2025 · score 28.8 · #185Mistral Small 3.1 24B · listed 17 Mar 2025 · score 31.0 · all versions #287DeepSeek V3 0324 · listed 24 Mar 2025 · score 55.4 · all versions #171Llama 4 Scout · listed 5 Apr 2025 · score 34.5 · #171Llama 4 Maverick · listed 5 Apr 2025 · score 36.3 · #167Llama 3 1 Nemotron Ultra 253B v1 · listed 7 Apr 2025 · score 42.7 · #139GPT-4.1 Nano · listed 14 Apr 2025 · score 35.3 · all versions #268GPT-4.1 Mini · listed 14 Apr 2025 · score 52.1 · all versions #190o4 Mini · listed 16 Apr 2025 · score 53.6 · #108Qwen3 235B A22B · listed 28 Apr 2025 · score 57.7 · #86Qwen3 32B · listed 28 Apr 2025 · score 43.2 · #138Qwen3 30B A3B · listed 28 Apr 2025 · score 37.7 · #165Mistral Medium 3 · listed 7 May 2025 · score 53.5 · all versions #185Claude Sonnet 4 · listed 22 May 2025 · score 57.7 · all versions #149Claude Opus 4 · listed 22 May 2025 · score 67.2 · all versions #99R1 0528 · listed 28 May 2025 · score 64.3 · #62gemma 3n E4B it · listed 3 Jun 2025 · score 32.9 · #174Gemini 2.5 Flash · listed 17 Jun 2025 · score 59.0 · all versions #140MiniMax M1 · listed 17 Jun 2025 · score 48.2 · all versions #205Mistral Small 3.2 24B · listed 20 Jun 2025 · score 45.1 · #135Qwen3 235B A22B Instruct 2507 · listed 21 Jul 2025 · score 64.7 · #60Qwen3 Coder 480B A35B · listed 23 Jul 2025 · score 54.3 · #105Qwen3 235B A22B Thinking 2507 · listed 25 Jul 2025 · score 57.2 · #90GLM 4.5 Air · listed 25 Jul 2025 · score 50.5 · #114GLM 4.5 · listed 25 Jul 2025 · score 60.9 · all versions #134Qwen3 30B A3B Instruct 2507 · listed 29 Jul 2025 · score 52.7 · #109gpt-oss-20b · listed 5 Aug 2025 · score 39.1 · #163gpt-oss-120b · listed 5 Aug 2025 · score 46.3 · #130GPT-5 Nano · listed 7 Aug 2025 · score 39.7 · all versions #258GPT-5 Mini · listed 7 Aug 2025 · score 54.5 · all versions #178GPT-5 · listed 7 Aug 2025 · score 68.6 · all versions #90GLM 4.5V · listed 11 Aug 2025 · score 44.1 · all versions #223Mistral Medium 3.1 · listed 13 Aug 2025 · score 58.5 · all versions #144DeepSeek V3.1 · listed 21 Aug 2025 · score 63.2 · all versions #117Kimi K2 0905 · listed 4 Sept 2025 · score 63.6 · all versions #115Qwen3 Next 80B A3B Instruct · listed 11 Sept 2025 · score 57.2 · #89Qwen3 Next 80B A3B Thinking · listed 11 Sept 2025 · score 49.3 · #117DeepSeek V3.1 Terminus · listed 22 Sept 2025 · score 62.4 · #70Qwen3 Max · listed 23 Sept 2025 · score 65.9 · all versions #106Qwen3 VL 235B A22B Instruct · listed 23 Sept 2025 · score 62.8 · #64Qwen3 VL 235B A22B Thinking · listed 23 Sept 2025 · score 56.4 · #95DeepSeek V3.2 Exp · listed 29 Sept 2025 · score 66.6 · all versions #101GLM 4.6 · listed 30 Sept 2025 · score 65.2 · all versions #109GPT-5 Pro · listed 6 Oct 2025 · score 50.3 · all versions #198Claude Haiku 4.5 · listed 15 Oct 2025 · score 62.6 · #65MiniMax M2 · listed 23 Oct 2025 · score 42.5 · all versions #236Kimi K2 Thinking · listed 6 Nov 2025 · score 68.5 · #47GPT-5.1 · listed 13 Nov 2025 · score 75.7 · all versions #60Olmo 3 32B Think · listed 19 Nov 2025 · score 32.4 · all versions #284DeepSeek V3.2 · listed 1 Dec 2025 · score 64.4 · #61Mistral Large 3 2512 · listed 1 Dec 2025 · score 62.2 · #71Nova 2 Lite · listed 2 Dec 2025 · score 40.4 · #155GLM 4.6V · listed 8 Dec 2025 · score 50.9 · #112Olmo 3.1 32B Think · listed 10 Dec 2025 · score 26.8 · #192GPT-5.2 · listed 10 Dec 2025 · score 70.6 · all versions #82GPT-5.2 Pro · listed 10 Dec 2025 · score 56.6 · all versions #161Nemotron 3 Nano 30B A3B · listed 14 Dec 2025 · score 33.4 · #173Gemini 3 Flash Preview · listed 17 Dec 2025 · score 80.8 · all versions #33GLM 4.7 · listed 22 Dec 2025 · score 72.0 · all versions #74MiniMax M2.1 · listed 23 Dec 2025 · score 53.3 · all versions #186GLM 4.7 Flash · listed 19 Jan 2026 · score 48.9 · all versions #202Solar Pro 3 · listed 27 Jan 2026 · score 52.7 · all versions #188Kimi K2.5 · listed 27 Jan 2026 · score 74.0 · all versions #67Step 3.5 Flash · listed 29 Jan 2026 · score 54.7 · all versions #175Qwen3 Coder Next · listed 4 Feb 2026 · score 54.0 · #107GLM 5 · listed 11 Feb 2026 · score 75.4 · all versions #62MiniMax M2.5 · listed 12 Feb 2026 · score 52.2 · all versions #189Qwen3.5 397B A17B · listed 16 Feb 2026 · score 70.4 · #42Claude Sonnet 4.6 · listed 17 Feb 2026 · score 79.8 · all versions #38Gemini 3.1 Pro Preview · listed 19 Feb 2026 · score 81.2 · #17Qwen3.5-Flash · listed 25 Feb 2026 · score 56.7 · all versions #160Qwen3.5-122B-A10B · listed 25 Feb 2026 · score 62.1 · #72Qwen3.5-27B · listed 25 Feb 2026 · score 59.6 · all versions #138Qwen3.5-35B-A3B · listed 25 Feb 2026 · score 56.2 · all versions #165Gemini 3.1 Flash Lite Preview · listed 3 Mar 2026 · score 65.7 · all versions #107Mercury 2 · listed 4 Mar 2026 · score 42.8 · all versions #231GPT-5.4 · listed 5 Mar 2026 · score 81.1 · all versions #30GPT-5.4 Pro · listed 5 Mar 2026 · score 63.4 · all versions #116Nemotron 3 Super · listed 11 Mar 2026 · score 46.6 · #128GLM 5 Turbo · listed 15 Mar 2026 · score 54.7 · #103Mistral Small 4 · listed 16 Mar 2026 · score 55.8 · #99GPT-5.4 Mini · listed 17 Mar 2026 · score 72.3 · #34GPT-5.4 Nano · listed 17 Mar 2026 · score 58.8 · #80MiniMax M2.7 · listed 18 Mar 2026 · score 62.7 · all versions #120Grok 4.20 · listed 31 Mar 2026 · score 81.0 · all versions #31Grok 4.20 Multi-Agent · listed 31 Mar 2026 · score 81.0 · #18Trinity Large Thinking · listed 1 Apr 2026 · score 50.4 · #115GLM 5V Turbo · listed 1 Apr 2026 · score 69.6 · #45Qwen3.6 Plus · listed 2 Apr 2026 · score 72.8 · all versions #70Gemma 4 31B · listed 2 Apr 2026 · score 74.1 · #29Gemma 4 26B A4B · listed 3 Apr 2026 · score 70.8 · #38granite 4.1 8b · listed 6 Apr 2026 · score 31.3 · all versions #286GLM 5.1 · listed 7 Apr 2026 · score 76.6 · all versions #56Kimi K2.6 · listed 20 Apr 2026 · score 76.5 · all versions #57MiMo-V2.5 · listed 22 Apr 2026 · score 70.3 · #43MiMo-V2.5-Pro · listed 22 Apr 2026 · score 77.1 · all versions #54Hy3 preview · listed 22 Apr 2026 · score 61.3 · all versions #132DeepSeek V4 Flash 0423 · listed 24 Apr 2026 · score 71.4 · all versions #78DeepSeek V4 Pro 0423 · listed 24 Apr 2026 · score 76.2 · all versions #59GPT-5.5 · listed 24 Apr 2026 · score 82.2 · #13GPT-5.5 Pro · listed 24 Apr 2026 · score 65.3 · #59Qwen3.6 27B · listed 27 Apr 2026 · score 63.8 · all versions #113Qwen3.6 Max Preview · listed 27 Apr 2026 · score 78.3 · all versions #50Qwen3.6 35B A3B · listed 27 Apr 2026 · score 62.1 · #73Mistral Medium 3.5 · listed 30 Apr 2026 · score 63.7 · #63Grok 4.3 · listed 30 Apr 2026 · score 71.4 · all versions #77Ring 2.6 1T · listed 14 May 2026 · score 60.6 · #77Gemini 3.5 Flash · listed 19 May 2026 · score 80.1 · all versions #36Qwen3.7 Max · listed 21 May 2026 · score 78.1 · all versions #51Claude Opus 4.8 · listed 27 May 2026 · score 83.1 · all versions #20MiniMax M3 · listed 31 May 2026 · score 71.3 · #37Qwen3.7 Plus · listed 3 Jun 2026 · score 73.7 · #30Kimi K2.7 Code · listed 12 Jun 2026 · score 67.7 · #50GLM 5.2 · listed 16 Jun 2026 · score 78.4 · all versions #47North Mini Code · listed 17 Jun 2026 · score 54.6 · #104Claude Sonnet 5 · listed 30 Jun 2026 · score 78.4 · #26Hy3 · listed 6 Jul 2026 · score 74.1 · all versions #65Grok 4.5 · listed 8 Jul 2026 · score 78.9 · all versions #44GPT-5.6 Sol · listed 9 Jul 2026 · score 84.1 · all versions #13GPT-5.6 Terra · listed 9 Jul 2026 · score 79.5 · #22GPT-5.6 Luna · listed 9 Jul 2026 · score 75.0 · all versions #63Intern S2 Preview 397B · listed 16 Jul 2026 · score 51.8 · #110Muse Spark 1.1 · listed 16 Jul 2026 · score 82.6 · all versions #22Kimi K3 · listed 16 Jul 2026 · score 83.5 · #12Inkling · listed 17 Jul 2026 · score 69.2 · #46Gemini 3.5 Flash Lite · listed 21 Jul 2026 · score 72.6 · #33Gemini 3.6 Flash · listed 21 Jul 2026 · score 80.4 · all versions #34Claude Opus 5 · listed 24 Jul 2026 · score 86.3 · all versions #9Inkling Small · listed 30 Jul 2026 · score 58.1 · #84DeepSeek V4 Flash 0731 · listed 31 Jul 2026 · score 78.7 · all versions #45Qwen3.8 Max (0803) · listed 3 Aug 2026 · score 83.2 · all versions #19Muse Spark 1.2 · listed 5 Aug 2026 · score 83.8 · all versions #17granite 4.2 3b · listed 7 Aug 2026 · score 36.4 · #166granite 4.2 30b · listed 7 Aug 2026 · score 42.0 · #145Muse Glimmer 30B · listed 9 Aug 2026 · score 71.5 · #36Solar Pro 4 · listed 10 Aug 2026 · score 50.7 · #113Nemotron 3.5 Lightning · listed 11 Aug 2026 · score 57.3 · #88Grok 4.6 · listed 12 Aug 2026 · score 77.3 · all versions #52DeepSeek V4 Pro 0813 · listed 12 Aug 2026 · score 78.5 · #25Gemini 3.7 Flash · listed 13 Aug 2026 · score 82.4 · all versions #23Dots3-Note Preview · listed 14 Aug 2026 · score 61.5 · #74Qwen3.8 27B · listed 14 Aug 2026 · score 71.5 · #35GLM 5.3 · listed 18 Aug 2026 · score 83.8 · #11GLM 5.3 Flash · listed 26 Aug 2026 · score 81.6 · #14Qwen3.8 Flash · listed 26 Aug 2026 · score 81.5 · #15Ling 3.0 Flash Fin · listed 27 Aug 2026 · score 66.0 · #58Hy4 preview · listed 28 Aug 2026 · score 76.9 · #28Granite 4.2 8B · listed 31 Aug 2026 · score 35.8 · #168Gemini 3.8 Flash · listed 2 Sept 2026 · score 84.1 · #9Muse Spark 1.3 · listed 2 Sept 2026 · score 85.6 · #8Qwen3.8 Max (0902) · listed 3 Sept 2026 · score 86.5 · #6GPT-6 Astra · listed 4 Sept 2026 · score 86.2 · #7MiniCPM5 2B · listed 6 Sept 2026 · score 56.9 · #92Mercury 2.5 · listed 8 Sept 2026 · score 56.8 · #93DeepSeek V4.1 Flash · listed 10 Sept 2026 · score 80.2 · #19Ling 3.0 Flash VL · listed 10 Sept 2026 · score 67.8 · #49Grok 4.7 · listed 21 Sept 2026 · score 87.4 · #4MiMo-V2.6-Pro · listed 21 Sept 2026 · score 87.3 · #5GPT-6 Sol · listed 22 Sept 2026 · score 88.4 · #2GPT-6 Luna · listed 22 Sept 2026 · score 78.4 · #27Command A+ · listed 22 Sept 2026 · score 57.5 · #87GPT-4 · listed 28 May 2023 · score 27.2 · all versions #299Mistral Large · listed 26 Feb 2024 · score 31.0 · all versions #288GPT-4 Turbo · listed 9 Apr 2024 · score 35.2 · #169GPT-4o (2024-05-13) · listed 13 May 2024 · score 41.7 · all versions #249o1 · listed 17 Dec 2024 · score 57.2 · all versions #154GPT-4.1 · listed 14 Apr 2025 · score 61.3 · all versions #131o3 · listed 16 Apr 2025 · score 67.4 · #52Gemini 2.5 Pro · listed 17 Jun 2025 · score 68.2 · all versions #92Claude Opus 4.1 · listed 5 Aug 2025 · score 76.5 · all versions #58Claude Sonnet 4.5 · listed 29 Sept 2025 · score 77.3 · all versions #53Claude Opus 4.5 · listed 24 Nov 2025 · score 81.2 · all versions #29GPT-5.2 Chat · listed 10 Dec 2025 · score 82.9 · all versions #21Claude Opus 4.6 · listed 4 Feb 2026 · score 86.3 · all versions #10Claude Opus 4.7 · listed 16 Apr 2026 · score 86.5 · all versions #7Claude Fable 5 · listed 9 Jun 2026 · score 87.7 · all versions #4Claude Fable 5.1 · listed 1 Sept 2026 · score 87.8 · #3Claude Opus 5.5 · listed 22 Sept 2026 · score 97.1 · #1Claude Opus 5.5Claude Fable 5.1Claude Opus 4.6GPT-5.2 ChatClaude Opus 4.5Claude Sonnet 4.5Claude Opus 4.1Gemini 2.5 Proo3GPT-4.1o1GPT-4o (2024-05-13)GPT-4 TurboMistral LargeGPT-4
Measured model, any version Best measured score among models listed by that date

Each dot is a measured model on the all-versions board, placed at the date it was first listed in a public catalogue, usually OpenRouter, which can trail the publisher’s announcement. Scores are today’s Index scores, so the line shows how today’s evidence ranks models by when they appeared, not what was known at the time. Superseded models stay in.

Models that set a new high

Measured models that set a new best score when listed, oldest first
First listedModelIndex scorePosition now
28 May 2023GPT-4OpenAI27.2All versions #299
26 Feb 2024Mistral LargeMistral AI31.0All versions #288
9 Apr 2024GPT-4 TurboOpenAI35.2#169
13 May 2024GPT-4o (2024-05-13)OpenAI41.7All versions #249
17 Dec 2024o1OpenAI57.2All versions #154
14 Apr 2025GPT-4.1OpenAI61.3All versions #131
16 Apr 2025o3OpenAI67.4#52
17 Jun 2025Gemini 2.5 ProGoogle68.2All versions #92
5 Aug 2025Claude Opus 4.1Anthropic76.5All versions #58
29 Sept 2025Claude Sonnet 4.5Anthropic77.3All versions #53
24 Nov 2025Claude Opus 4.5Anthropic81.2All versions #29
10 Dec 2025GPT-5.2 ChatOpenAI82.9All versions #21
4 Feb 2026Claude Opus 4.6Anthropic86.3All versions #10
16 Apr 2026Claude Opus 4.7Anthropic86.5All versions #7
9 Jun 2026Claude Fable 5Anthropic87.7All versions #4
1 Sept 2026Claude Fable 5.1Anthropic87.8#3
22 Sept 2026Claude Opus 5.5Anthropic97.1Preliminary#1

How to read these charts

Explore further