Model intelligence
236 AI models
Capability (Epoch Capabilities Index and individual benchmarks from Epoch AI), API list prices and context windows (OpenRouter), release dates and open-weight status. Benchmarks are shown individually - never merged into one invented score.
Epoch Capabilities Index by release date
Capability frontier
Source: Epoch AI (CC BY 4.0). ECI is Epoch's item-response model over many benchmarks; each model page shows its 90% confidence interval.
Blended $/1M tokens (3:1 input:output) vs ECI
Price vs capability
Only models with both an ECI score and an OpenRouter price. Normalisation: 75% input + 25% output price. Click a point to open the model.
| # | Model | ECI ↑ | Context | In $/M | Weights | Released |
|---|---|---|---|---|---|---|
| 101 | GPT-5.1 OpenAI· reasoning | 149.7 | 400K | $1.25 | Closed | 13 Nov 2025 |
| 102 | GPT-5 OpenAI· reasoning | 150.0 | 400K | $1.25 | Closed | 7 Aug 2025 |
| 103 | Claude Opus 4.5 Anthropic· reasoning | 150.1 | 200K | $5.00 | Closed | 24 Nov 2025 |
| 104 | GPT-5 Pro OpenAI· reasoning | 150.3 | 400K | $15.0 | Closed | 7 Oct 2025 |
| 105 | Gemini 3 Flash | 151.8 | — | — | Closed | 17 Dec 2025 |
| 106 | Grok 4.20 xAI· reasoning | 152.0 | 2M | $1.25 | Closed | 17 Feb 2026 |
| 107 | Muse Spark | 152.1 | — | — | Closed | 8 Apr 2026 |
| 108 | Claude Sonnet 4.6 Anthropic· reasoning | 152.3 | 1M | $3.00 | Closed | 17 Feb 2026 |
| 109 | Gemini 3 Pro | 152.9 | — | — | Closed | 18 Nov 2025 |
| 110 | GPT-5.2 OpenAI· reasoning | 153.5 | 400K | $1.75 | Closed | 11 Dec 2025 |
| 111 | Qwen3.7 Max Alibaba (Qwen)· reasoning | 153.7 | 1M | $1.48 | Closed | 19 May 2026 |
| 112 | Grok 4.5 xAI· reasoning | 154.0 | 500K | $2.00 | Closed | 8 Jul 2026 |
| 113 | Muse Spark 1.1 Meta· reasoning | 154.3 | 1.05M | $1.25 | Closed | 9 Jul 2026 |
| 114 | Gemini 3.6 Flash Google· reasoning | 154.3 | 1.05M | $0.75 | Closed | 21 Jul 2026 |
| 115 | Gemini 3.5 Flash Google· reasoning | 154.5 | 1.05M | $1.50 | Closed | 19 May 2026 |
| 116 | Gemini 3.1 Pro | 154.8 | — | — | Closed | 19 Feb 2026 |
| 117 | Muse Spark 1.2 Meta· reasoning | 155.0 | 1.05M | $1.25 | Closed | 5 Aug 2026 |
| 118 | Qwen3.8 Max (0902) Alibaba (Qwen)· reasoning | 155.2 | 1M | $2.00 | Closed | 1 Sep 2026 |
| 119 | Claude Opus 4.6 Anthropic· reasoning | 155.3 | 1M | $5.00 | Closed | 5 Feb 2026 |
| 120 | GPT-5.2 Pro OpenAI· reasoning | 155.4 | 400K | $21.0 | Closed | 11 Dec 2025 |
| 121 | Claude Sonnet 5 Anthropic· reasoning | 156.3 | 1M | $2.00 | Closed | 30 Jun 2026 |
| 122 | Claude Opus 4.7 Anthropic· reasoning | 156.3 | 1M | $5.00 | Closed | 16 Apr 2026 |
| 123 | GPT-5.6 Luna OpenAI· reasoning | 156.4 | 1.05M | $0.20 | Closed | 9 Jul 2026 |
| 124 | Qwen 3.8 Max | 156.6 | — | — | Closed | 2 Aug 2026 |
| 125 | Grok 4.6 xAI· reasoning | 156.6 | 500K | $2.00 | Closed | 12 Aug 2026 |
| 126 | GPT-5.3-Codex OpenAI· reasoning | 156.8 | 400K | $1.75 | Closed | 5 Feb 2026 |
| 127 | Muse Spark 1.3 Meta· reasoning | 156.9 | 1.05M | $1.25 | Closed | 2 Sep 2026 |
| 128 | GPT-5.4 OpenAI· reasoning | 156.9 | 1.05M | $2.50 | Closed | 5 Mar 2026 |
| 129 | Gemini 3.8 Flash Google· reasoning | 156.9 | 1.05M | $0.75 | Closed | 2 Sep 2026 |
| 130 | Gemini 3.7 Flash Google· reasoning | 157.4 | 1.05M | $0.75 | Closed | 13 Aug 2026 |
| 131 | Claude Opus 4.8 Anthropic· reasoning | 158.3 | 1M | $5.00 | Closed | 28 May 2026 |
| 132 | GPT-5.4 Pro OpenAI· reasoning | 159.1 | 1.05M | $30.0 | Closed | 5 Mar 2026 |
| 133 | GPT-5.5 OpenAI· reasoning | 159.2 | 1.05M | $5.00 | Closed | 23 Apr 2026 |
| 134 | GPT-5.6 Terra OpenAI· reasoning | 159.8 | 1.05M | $2.00 | Closed | 9 Jul 2026 |
| 135 | GPT-5.6 Sol OpenAI· reasoning | 161.8 | 1.05M | $2.00 | Closed | 9 Jul 2026 |
| 136 | Claude Fable 5 Anthropic· reasoning | 162.2 | 1M | $10.0 | Closed | 9 Jun 2026 |
| 137 | GPT-5.5 Pro OpenAI· reasoning | 162.4 | 1.05M | $30.0 | Closed | 23 Apr 2026 |
| 138 | Claude Opus 5 Anthropic· reasoning | 162.9 | 1M | $5.00 | Closed | 24 Jul 2026 |
| 139 | Claude Fable 5.1 Anthropic· reasoning | 164.8 | 1M | $10.0 | Closed | 1 Sep 2026 |
| 140 | Claude Sonnet 5.5 Anthropic· reasoning | 165.2 | 1M | $2.00 | Closed | 28 Sep 2026 |
| 141 | GPT-6 Astra OpenAI· reasoning | 166.5 | 1.05M | $10.0 | Closed | 3 Sep 2026 |
| 142 | Claude Opus 5.5 Anthropic· reasoning | 167.3 | 1M | $4.00 | Closed | 22 Sep 2026 |
| 143 | GPT-6.1 Sol (unknown thinking) | — | — | — | Closed | 29 Sep 2026 |
| 144 | GPT-6.1 Sol (medium) | — | — | — | Closed | 29 Sep 2026 |
| 145 | GPT-6 Sol (xhigh) | — | — | — | Closed | 22 Sep 2026 |
| 146 | GPT-6 Sol (medium) | — | — | — | Closed | 22 Sep 2026 |
| 147 | GPT-6 Luna (xhigh) | — | — | — | Closed | 22 Sep 2026 |
| 148 | GPT-6 Sol (low) | — | — | — | Closed | 22 Sep 2026 |
| 149 | GPT-6 Luna (medium) | — | — | — | Closed | 22 Sep 2026 |
| 150 | GPT-6 Luna (low) | — | — | — | Closed | 22 Sep 2026 |
| 151 | GPT-6 Sol (none) | — | — | — | Closed | 22 Sep 2026 |
| 152 | GPT-6 Luna (none) | — | — | — | Closed | 22 Sep 2026 |
| 153 | GPT-6 Sol (unknown thinking) | — | — | — | Closed | 22 Sep 2026 |
| 154 | GPT-6 Luna (unknown thinking) | — | — | — | Closed | 22 Sep 2026 |
| 155 | Grok 4.7 (unknown) | — | — | — | Closed | 21 Sep 2026 |
| 156 | Grok 4.7 (xhigh) | — | — | — | Closed | 21 Sep 2026 |
| 157 | SWE-2 (max) | — | — | — | Closed | 10 Sep 2026 |
| 158 | Mercury 2.5 (unknown) | — | — | — | Closed | 8 Sep 2026 |
| 159 | GPT-6 Astra (pro, max) | — | — | — | Closed | 3 Sep 2026 |
| 160 | GLM-5.3 (unknown thinking) | — | — | — | Closed | 14 Aug 2026 |
| 161 | Qwen3.8 Max (xhigh) | — | — | — | Closed | 2 Aug 2026 |
| 162 | Qwen3.8 Max (unknown) | — | — | — | Closed | 2 Aug 2026 |
| 163 | Gemini 3.5 Flash-Lite (medium) | — | — | — | Closed | 21 Jul 2026 |
| 164 | GPT-5.6 Sol (pro, max) | — | — | — | Closed | 9 Jul 2026 |
| 165 | GPT-5.6 Sol (pro, xhigh) | — | — | — | Closed | 9 Jul 2026 |
| 166 | GPT-5.6 Sol (pro, unknown thinking) | — | — | — | Closed | 9 Jul 2026 |
| 167 | Composer 2.5 | — | — | — | Closed | 18 May 2026 |
| 168 | AI co-mathematician | — | — | — | Closed | 8 May 2026 |
| 169 | Claude Mythos Preview (Early) | — | — | — | Closed | 7 Apr 2026 |
| 170 | SWE-1.6 | — | — | — | Closed | 7 Apr 2026 |
| 171 | Qwen 3.6 Plus (2026-04-02) | — | — | — | Closed | 31 Mar 2026 |
| 172 | GPT-5.4 mini (xhigh) | — | — | — | Closed | 17 Mar 2026 |
| 173 | GPT-5.4 mini (none) | — | — | — | Closed | 17 Mar 2026 |
| 174 | GPT-5.4 nano (no thinking) | — | — | — | Closed | 17 Mar 2026 |
| 175 | GPT-5.4 nano (low) | — | — | — | Closed | 17 Mar 2026 |
| 176 | GPT-5.4 mini (low) | — | — | — | Closed | 17 Mar 2026 |
| 177 | GPT-5.4 mini (medium) | — | — | — | Closed | 17 Mar 2026 |
| 178 | GPT-5.4 nano (medium) | — | — | — | Closed | 17 Mar 2026 |
| 179 | GPT-5.4 nano (xhigh) | — | — | — | Closed | 17 Mar 2026 |
| 180 | GPT-5.4 mini (unknown thinking) | — | — | — | Closed | 17 Mar 2026 |
| 181 | GPT-5.4 Pro (web) | — | — | — | Closed | 5 Mar 2026 |
| 182 | Gemini 3 Deep Think | — | — | — | Closed | 12 Feb 2026 |
| 183 | GPT-5.3 Codex (xhigh) | — | — | — | Closed | 5 Feb 2026 |
| 184 | GPT-5.2 Pro (web) | — | — | — | Closed | 11 Dec 2025 |
| 185 | Grok 4.1 Fast | — | — | — | Closed | 19 Nov 2025 |
| 186 | Gemini 3 Pro Preview | — | — | — | Closed | 18 Nov 2025 |
| 187 | Grok 4.1 | — | — | — | Closed | 17 Nov 2025 |
| 188 | Gemini 2.5 Flash-Lite (Sep 2025) | — | — | — | Closed | 25 Sep 2025 |
| 189 | Qwen3-Max-Instruct | — | — | — | Closed | 24 Sep 2025 |
| 190 | GPT-5-codex | — | — | — | Closed | 15 Sep 2025 |
| 191 | GPT-5 mini (minimal) | — | — | — | Closed | 7 Aug 2025 |
| 192 | GPT-5 nano (low) | — | — | — | Closed | 7 Aug 2025 |
| 193 | GPT-5 nano (minimal) | — | — | — | Closed | 7 Aug 2025 |
| 194 | GPT-5 mini (medium) | — | — | — | Closed | 7 Aug 2025 |
| 195 | GPT-5 nano (medium) | — | — | — | Closed | 7 Aug 2025 |
| 196 | GPT-5 mini (low) | — | — | — | Closed | 7 Aug 2025 |
| 197 | GPT-5 mini (unknown thinking) | — | — | — | Closed | 7 Aug 2025 |
| 198 | GPT-5 nano (unknown thinking) | — | — | — | Closed | 7 Aug 2025 |
| 199 | Gemini 2.5 Deep Think | — | — | — | Closed | 1 Aug 2025 |
| 200 | Grok 4 Heavy (web app) | — | — | — | Closed | 10 Jul 2025 |
Showing 100 of 236. Benchmark cells are % (best reported setting). 3.9K benchmark results in total.← PrevNext →Export CSV