Epoch Capabilities Index Leaderboard
composite189 models · Updated November 2026
On the Epoch Capabilities Index benchmark, GPT-6 Astra ranks #1 with a score of 169, while Mistral Nemo offers the best score-per-dollar at $0.03/1M output tokens. The full ranking, with cost per million tokens, is below.
💰 Best value
Mistral Nemo — score 119 at $0.03/1M output tokens
| # | Model | Score | Output / 1M | Context |
|---|---|---|---|---|
| 1 | GPT-6 AstraOpenAI | 169 | $50.00 | 1.1M |
| 2 | Claude Opus 5.5Anthropic | 168 | $20.00 | 1M |
| 3 | GPT-6.1 TerraOpenAI | 168 | $5.00 | 1.1M |
| 4 | GPT-6.1 SolOpenAI | 168 | $10.00 | 1.1M |
| 5 | Claude Sonnet 5.5Anthropic | 167 | $10.00 | 1M |
| 6 | Claude Fable 5.1Anthropic | 166 | $50.00 | 1M |
| 7 | Gemini 4 ArgonGoogle | 166 | $10.00 | 1M |
| 8 | Claude Opus 5Anthropic | 164 | $25.00 | 1M |
| 9 | GPT-6 SolOpenAI | 164 | $10.00 | 1.1M |
| 10 | GPT-5.5 ProOpenAI | 163 | $180.00 | 1.1M |
| 11 | Claude Fable 5Anthropic | 163 | $50.00 | 1M |
| 12 | GPT-5.6 SolOpenAI | 163 | $20.00 | 1.1M |
| 13 | Muse WonderMeta | 162 | $20.00 | 1M |
| 14 | Claude Haiku 5.5Anthropic | 160 | $0.50 | 1M |
| 15 | GPT-5.6 TerraOpenAI | 160 | $12.00 | 1.1M |
| 16 | GPT-5.4 ProOpenAI | 160 | $180.00 | 1.1M |
| 17 | Claude Opus 4.8Anthropic | 160 | $25.00 | 1M |
| 18 | GPT-5.5OpenAI | 159 | $30.00 | 1.1M |
| 19 | DeepSeek V4.1 FlashDeepSeek | 159 | $0.50 | 1M |
| 20 | Kimi K3MoonshotAI | 158 | $15.00 | 1M |
| 21 | Gemini 3.7 FlashGoogle | 158 | $3.75 | 1M |
| 22 | Qwen3.8 MaxAlibaba | 158 | $6.00 | 1M |
| 23 | Gemini 3.8 FlashGoogle | 158 | $3.75 | 1M |
| 24 | Muse Spark 1.3Meta | 157 | $4.25 | 1M |
| 25 | GPT-5.4OpenAI | 157 | $15.00 | 1.1M |
| 26 | Claude Sonnet 5Anthropic | 157 | $10.00 | 1M |
| 27 | Grok 4.6SpaceXAI | 157 | $6.00 | 500K |
| 28 | GPT-6 LunaOpenAI | 157 | $0.50 | 1.1M |
| 29 | Claude Opus 4.7Anthropic | 157 | $25.00 | 1M |
| 30 | GPT-5.6 LunaOpenAI | 157 | $1.20 | 1.1M |
| 31 | GPT-5.2 ProOpenAI | 157 | $168.00 | 400K |
| 32 | GLM 5.3Z.ai | 156 | $1.40 | 1M |
| 33 | DeepSeek V4 Pro 0423DeepSeek | 156 | $3.20 | 1M |
| 34 | Claude Opus 4.6Anthropic | 156 | $25.00 | 1M |
| 35 | Gemini-3.1-ProPoe | 155 | $12.12 | 1M |
| 36 | Muse Spark 1.2Meta | 155 | $4.25 | 1M |
| 37 | Muse Spark OpenMeta | 155 | $4.25 | 1M |
| 38 | Gemini 3.5 FlashGoogle | 155 | $9.00 | 1M |
| 39 | DeepSeek V4 Flash 0423DeepSeek | 155 | $0.18 | 1M |
| 40 | Gemini 3.6 FlashGoogle | 155 | $3.75 | 1M |
| 41 | Grok 4.7SpaceXAI | 155 | $6.00 | 500K |
| 42 | Grok 4.5SpaceXAI | 154 | $6.00 | 500K |
| 43 | Qwen3.7 MaxQwen | 154 | $4.43 | 1M |
| 44 | Muse Spark 1.1Meta | 154 | $4.25 | 1M |
| 45 | GPT-5.2OpenAI | 154 | $14.00 | 400K |
| 46 | GLM 5.3 FlashZ.ai | 153 | $0.17 | 1M |
| 47 | Gemini 3 ProOpenCode | 153 | $12.00 | 1M |
| 48 | Claude Sonnet 4.6Anthropic | 153 | $15.00 | 1M |
| 49 | Grok 4.20SpaceXAI | 152 | $2.50 | 2M |
| 50 | Muse SparkMeta AI | 152 | — | — |
| 51 | GLM 5.2Z.ai | 152 | $6.80 | 1M |
| 52 | Gemini 3 FlashOpenCode | 152 | $3.00 | 1M |
| 53 | Qwen3.8 27BQwen | 152 | $1.88 | 1M |
| 54 | Kimi K2.6MoonshotAI | 151 | $3.41 | 262K |
| 55 | GPT-5 ProOpenAI | 151 | $120.00 | 400K |
| 56 | Mistral Large 4Mistral | 151 | $2.10 | 1M |
| 57 | Inkling SmallThinking Machines | 150 | $1.20 | 524K |
| 58 | Beam 1.0Vercel AI Gateway | 150 | $0.00 | 1M |
| 59 | Kimi K2.7 CodeMoonshotAI | 150 | $3.50 | 262K |
| 60 | Claude Opus 4.5Anthropic | 150 | $25.00 | 200K |
| 61 | GPT-5OpenAI | 150 | $10.00 | 400K |
| 62 | GPT-5.1OpenAI | 150 | $10.00 | 400K |
| 63 | GLM 5.1Z.ai | 150 | $4.40 | 205K |
| 64 | GPT-5.4 MiniOpenAI | 149 | $4.50 | 400K |
| 65 | InklingThinking Machines | 149 | $4.05 | 524K |
| 66 | Kimi K2.5MoonshotAI | 148 | $2.50 | 262K |
| 67 | Qwen3.6 PlusQwen | 148 | $1.95 | 1M |
| 68 | o3 ProOpenAI | 147 | $80.00 | 200K |
| 69 | Qwen3.7 PlusAlibaba | 147 | $1.60 | 1M |
| 70 | MiniMax M3MiniMax | 147 | $1.20 | 1M |
| 71 | o3OpenAI | 147 | $8.00 | 200K |
| 72 | Qwen3.5 397B-A17BAlibaba | 147 | $3.60 | 262K |
| 73 | Claude Sonnet 4.5Anthropic | 147 | $15.00 | 1M |
| 74 | MiniMax M2.5MiniMax | 147 | $0.95 | 205K |
| 75 | Grok-4Poe | 146 | $15.00 | 256K |
| 76 | Qwen3.6 27BQwen | 146 | $2.70 | 262K |
| 77 | Qwen3.5 PlusAlibaba | 146 | $2.40 | 1M |
| 78 | GPT-5.4 NanoOpenAI | 146 | $1.25 | 400K |
| 79 | DeepSeek V3.2DeepSeek | 146 | $0.38 | 164K |
| 80 | Nemotron 3 UltraNVIDIA | 146 | $2.40 | 262K |
| 81 | MiniMax M2.7MiniMax | 146 | $1.20 | 205K |
| 82 | Kimi K2 0711MoonshotAI | 146 | $2.30 | 131K |
| 83 | GLM 5Z.ai | 146 | $3.20 | 205K |
| 84 | o4 MiniOpenAI | 146 | $4.40 | 200K |
| 85 | GPT-5 MiniOpenAI | 145 | $2.00 | 400K |
| 86 | Gemini 2.5 Pro (Jun 2025)Google DeepMind | 145 | — | — |
| 87 | DeepSeek V3.2 ExpDeepSeek | 145 | $0.41 | 164K |
| 88 | Gemini 3.5 Flash LiteGoogle | 145 | $2.50 | 1M |
| 89 | Gemini 3.1 Flash LiteGoogle | 144 | $1.50 | 1M |
| 90 | Grok 4 FastZenMux | 144 | $0.50 | 2M |
| 91 | Qwen3.7 FlashAlibaba | 144 | $0.13 | 1M |
| 92 | Claude Opus 4.1Anthropic | 144 | $75.00 | 200K |
| 93 | Qwen3.6 35B-A3BAlibaba | 144 | $2.25 | 262K |
| 94 | Qwen3 235B-A22BAlibaba | 144 | $2.80 | 131K |
| 95 | GLM 4.7Z.ai | 144 | $2.20 | 205K |
| 96 | Qwen3.5 FlashAlibaba | 144 | $0.40 | 1M |
| 97 | Gemma 4 31B ITQVAC | 143 | $0.00 | 262K |
| 98 | Qwen3.6 FlashQwen | 143 | $1.13 | 1M |
| 99 | Gemini 2.5 FlashGoogle | 143 | $2.50 | 1M |
| 100 | Qwen3.5 35B-A3BAlibaba | 143 | $2.00 | 262K |
| 101 | GPT-5.5 InstantZenMux | 142 | $30.00 | 400K |
| 102 | Claude Opus 4Anthropic | 142 | $75.00 | 200K |
| 103 | Qwen3 MaxAlibaba | 142 | $6.00 | 262K |
| 104 | Claude Haiku 4.5Anthropic | 142 | $5.00 | 200K |
| 105 | Gemma 4 26B A4BGoogle | 142 | $0.23 | 262K |
| 106 | o1OpenAI | 142 | $60.00 | 200K |
| 107 | Claude Sonnet 4 (latest)Anthropic | 142 | $15.00 | 200K |
| 108 | Mistral Medium 3.5Mistral | 141 | $7.50 | 262K |
| 109 | DeepSeek-R1Azure | 141 | $5.40 | 164K |
| 110 | Claude 3.7 SonnetHelicone | 141 | $15.00 | 200K |
| 111 | o3 MiniOpenAI | 140 | $4.40 | 200K |
| 112 | Grok 3 MiniPoe | 140 | $0.50 | 131K |
| 113 | gpt-oss-120b (batch)OpenAI | 140 | $0.14 | 131K |
| 114 | DeepSeek V3.1DeepSeek | 140 | $0.95 | 164K |
| 115 | Qwen3 30B A3BQwen | 140 | $0.50 | 41K |
| 116 | Qwen3.5-9BQwen | 139 | $0.25 | 262K |
| 117 | GPT-5 NanoOpenAI | 139 | $0.40 | 400K |
| 118 | Qwen3 235B A22B Instruct 2507Qwen | 139 | $0.55 | 262K |
| 119 | Qwen3 32BQwen | 138 | $0.28 | 131K |
| 120 | Grok 3Poe | 138 | $15.00 | 131K |
| 121 | Qwen3 14BAlibaba | 138 | $1.40 | 131K |
| 122 | gpt-oss-20b (batch)OpenAI | 138 | $0.11 | 131K |
| 123 | QwQ 32BAlibaba (China) | 138 | $0.86 | 131K |
| 124 | Qwen3 30B A3B Instruct 2507Qwen | 137 | $0.30 | 262K |
| 125 | GPT-4.1OpenAI | 137 | $8.00 | 1M |
| 126 | GPT-4.5OpenAI | 137 | — | — |
| 127 | Qwen3 8BAlibaba | 136 | $0.70 | 131K |
| 128 | DeepSeek V3DeepSeek | 136 | $1.03 | 164K |
| 129 | o1-miniAzure | 136 | $4.40 | 128K |
| 130 | Gemini 2.0 FlashGoogle | 135 | $0.40 | 1M |
| 131 | Gemini 2.0 ProGoogle DeepMind | 135 | — | — |
| 132 | GPT-4.1 MiniOpenAI | 135 | $1.60 | 1M |
| 133 | o1-previewOpenAI | 135 | — | — |
| 134 | Mistral Medium 3Mistral | 134 | $2.00 | 131K |
| 135 | Gemini 2.5 Flash LiteGoogle | 134 | $0.40 | 1M |
| 136 | Claude 3.5 SonnetDigitalOcean | 133 | $15.00 | 200K |
| 137 | Qwen 2.5 MaxNanoGPT | 133 | $6.39 | 32K |
| 138 | Llama 4 MaverickMeta | 132 | $0.65 | 524K |
| 139 | Gemini 1.5 ProGoogle DeepMind | 132 | — | — |
| 140 | Phi 4Microsoft | 131 | $0.14 | 16K |
| 141 | Grok-2xAI | 131 | — | — |
| 142 | Gemma 3 27BGoogle | 130 | $0.35 | 131K |
| 143 | Llama 4 ScoutMeta | 130 | $0.30 | 1.3M |
| 144 | GPT-4.1 NanoOpenAI | 130 | $0.40 | 1M |
| 145 | Gemini 1.5 FlashNanoGPT | 129 | $0.31 | 2M |
| 146 | GPT-4oOpenAI | 129 | $10.00 | 128K |
| 147 | Qwen2.5 72BNanoGPT | 129 | $0.40 | 131K |
| 148 | Llama 3.1-405BMeta AI | 129 | — | — |
| 149 | Mistral Large 2Mistral AI | 129 | — | — |
| 150 | Mistral Small 3.1Azure | 128 | $0.30 | 128K |
| 151 | Llama 3.3 70BVenice AI | 127 | $2.80 | 128K |
| 152 | GPT-4 TurboOpenAI | 127 | $30.00 | 128K |
| 153 | Claude 3.5 HaikuHelicone | 127 | $4.00 | 200K |
| 154 | Claude 3 OpusLLM Gateway | 127 | $75.00 | 200K |
| 155 | GPT-4o-miniOpenAI | 126 | $0.60 | 128K |
| 156 | Llama 3.1 70BWeights & Biases | 126 | $0.80 | 128K |
| 157 | GPT-4OpenAI | 126 | $60.00 | 8K |
| 158 | Qwen2-72BAlibaba | 125 | — | — |
| 159 | Gemma 3 12BGoogle | 124 | $0.15 | 131K |
| 160 | Llama 3-70BMeta AI | 123 | — | — |
| 161 | Gemma 2 27BGoogle | 122 | $0.65 | 8K |
| 162 | Mistral LargeMistral AI | 122 | $6.00 | 128K |
| 163 | Mixtral 8x22BMistral | 122 | $6.00 | 64K |
| 164 | Claude 3 SonnetAnthropic | 121 | — | — |
| 165 | Claude 2Anthropic | 120 | — | — |
| 166 | Qwen 2.5 Coder 32BAbacus | 119 | $0.79 | 128K |
| 167 | Claude 2.1Anthropic | 119 | — | — |
| 168 | Command R (08-2024)Cohere | 119 | $0.60 | 128K |
| 169 | Claude 3 HaikuAnthropic | 119 | $1.25 | 200K |
| 170 | Mistral NemoMistral | 119 | $0.03 | 131K |
| 171 | GPT-3.5 TurboOpenAI | 119 | $1.50 | 16K |
| 172 | Mixtral 8x7BMistral | 118 | $0.70 | 32K |
| 173 | Ministral 3B (latest)Mistral | 118 | $0.10 | 131K |
| 174 | Yi-34B01.AI | 117 | — | — |
| 175 | phi-3-mini 3.8BMicrosoft | 117 | — | — |
| 176 | Gemini 1.0 ProGoogle DeepMind | 117 | — | — |
| 177 | Llama 3-8BMeta AI | 116 | — | — |
| 178 | Llama 3.1 8BGroq | 116 | $0.08 | 131K |
| 179 | Gemma 3 4BGoogle | 116 | $0.10 | 131K |
| 180 | Llama 2-70BMeta AI | 114 | — | — |
| 181 | Falcon-180BTechnology Innovation Institute | 112 | — | — |
| 182 | LLaMA-65BMeta AI | 110 | — | — |
| 183 | Llama 2-13BMeta AI | 106 | — | — |
| 184 | Llama 2-34BMeta AI | 105 | — | — |
| 185 | Falcon-40BTechnology Innovation Institute | 104 | — | — |
| 186 | LLaMA-13BMeta AI | 101 | — | — |
| 187 | Llama 2-7BMeta AI | 98.8 | — | — |
| 188 | LLaMA-7BMeta AI | 96.9 | — | — |
| 189 | Qwen2.5-Coder-0.5BPioneer | 88.0 | $0.10 | 33K |
Frequently asked questions
What is the Epoch Capabilities Index benchmark?
Epoch Capabilities Index is a composite used to evaluate and rank large language models. This page lists 189 models by their Epoch Capabilities Index score.
Which model is best on Epoch Capabilities Index?
GPT-6 Astra (OpenAI) currently leads the Epoch Capabilities Index leaderboard with a score of 169.
What is the best value model on Epoch Capabilities Index?
Mistral Nemo offers the best score-per-dollar on Epoch Capabilities Index — a score of 119 at $0.03 per 1M output tokens.
Pricing is indicative — confirm with the provider before production use. Updated November 2026.