Epoch Capabilities Index
Combines scores from many benchmarks into one general capability scale, so models can be compared even after single benchmarks saturate.
- Unit
- ECI
- Results
- 270 results160 Tokscale models matched, 270 in the source
- Last update
- Oct 15, 2026Newest model release
- Retrieved
- Oct 21, 2026Source data through Oct 15, 2026
Conditions. One score per model, with the confidence interval Epoch AI publishes. Epoch fits the index jointly across all results, so a model's score can move when new results are added.
Attribution. Epoch AI, ‘Capabilities & benchmarking’, epoch.ai/benchmarks (CC BY 4.0)
Results
One row per model: its best reported result. The rank is the position in this list, by score; the source's own ranking can differ.
Tokscale usage counts accounts that used the model in public Tokscale submissions, Oct 14 – Nov 12, 2026. It is shown next to the score and does not explain it.
| Rank | Model | Score (ECI) | Date | Tokscale usage (30 days) | Source link |
|---|---|---|---|---|---|
| 1 | 168.8Confidence interval 165.3–173.9 | Sep 22, 2026Model released | 709active users | Source for Claude Opus 5.5 (opens in a new tab) | |
| 2 | 168.8Confidence interval 165.4–174.5 | Sep 3, 2026Model released | 563active users | Source for GPT-6 Astra (opens in a new tab) | |
| 3 | 168.4Confidence interval 165.0–174.4 | Sep 29, 2026Model released | 660active users | Source for GPT-6.1 Sol (opens in a new tab) | |
| 4 | 166.5Confidence interval 163.5–171.3 | Sep 1, 2026Model released | 361active users | Source for Claude Fable 5.1 (opens in a new tab) | |
| 5 | 166.4Confidence interval 163.2–171.5 | Sep 30, 2026Model released | 82active users | Source for Gemini 4 Argon (opens in a new tab) | |
| 6 | 166.2Confidence interval 163.1–171.3 | Sep 28, 2026Model released | 555active users | Source for Claude Sonnet 5.5 (opens in a new tab) | |
| 7 | 163.9Confidence interval 161.1–168.2 | Jul 24, 2026Model released | 163active users | Source for Claude Opus 5 (opens in a new tab) | |
| 8 | 163.7Confidence interval 160.9–167.4 | Sep 22, 2026Model released | 246active users | Source for GPT-6 Sol (opens in a new tab) | |
| 9 | 163.4Confidence interval 160.5–168.2 | Jun 9, 2026Model released | 41active users | Source for Claude Fable 5 (opens in a new tab) | |
| 10 | 162.9Confidence interval 160.0–166.6 | Apr 23, 2026Model released | 1active users | Source for GPT-5.5 Pro (opens in a new tab) | |
| 11 | 162.8Confidence interval 160.2–167.1 | Jul 9, 2026Model released | 217active users | Source for GPT-5.6 Sol (opens in a new tab) | |
| 12 | Muse Wonder 1.0Listed as Muse Wonder | 161.8Confidence interval 159.4–165.4 | Oct 12, 2026Model released | 17active users | Source for Muse Wonder 1.0 (opens in a new tab) |
| 13 | 160.1Confidence interval 157.3–163.3 | Jul 9, 2026Model released | 120active users | Source for GPT-5.6 Terra (opens in a new tab) | |
| 14 | 159.7Confidence interval 157.3–163.7 | Apr 23, 2026Model released | 30active users | Source for GPT-5.5 (opens in a new tab) | |
| 15 | 159.6Confidence interval 157.1–162.7 | Mar 5, 2026Model released | No Tokscale usage in the last 30 days | Source for GPT-5.4 Pro (opens in a new tab) | |
| 16 | 159.3Confidence interval 156.2–162.5 | Oct 7, 2026Model released | 477active users | Source for Claude Haiku 5.5 (opens in a new tab) | |
| 17 | 159.0Confidence interval 156.9–161.9 | May 28, 2026Model released | 222active users | Source for Claude Opus 4.8 (opens in a new tab) | |
| 18 | 158.0Confidence interval 155.9–161.2 | Aug 13, 2026Model released | 4active users | Source for Gemini 3.7 Flash (opens in a new tab) | |
| 19 | 157.9Confidence interval 155.1–160.6 | Jul 16, 2026Model released | 56active users | Source for Kimi K3 (opens in a new tab) | |
| 20 | 157.4Confidence interval 155.1–161.1 | Sep 2, 2026Model released | 66active users | Source for Gemini 3.8 Flash (opens in a new tab) | |
| 21 | Qwen3.8 MaxListed as Qwen 3.8 Max | 157.4Confidence interval 155.3–160.3 | Aug 2, 2026Model released | 16active users | Source for Qwen3.8 Max (opens in a new tab) |
| 22 | 157.3Confidence interval 155.4–159.8 | Mar 5, 2026Model released | 1active users | Source for GPT-5.4 (opens in a new tab) | |
| 23 | 157.2Confidence interval 155.3–159.8 | Sep 2, 2026Model released | 21active users | Source for Muse Spark 1.3 (opens in a new tab) | |
| 24 | 157.1Confidence interval 153.6–161.4 | Feb 5, 2026Model released | 5active users | Source for GPT-5.3 Codex (opens in a new tab) | |
| 25 | 156.9Confidence interval 154.1–159.7 | Jun 30, 2026Model released | 138active users | Source for Claude Sonnet 5 (opens in a new tab) | |
| 26 | 156.8Confidence interval 154.9–159.2 | Apr 16, 2026Model released | 27active users | Source for Claude Opus 4.7 (opens in a new tab) | |
| 27 | 156.8Confidence interval 154.3–159.4 | Sep 22, 2026Model released | 433active users | Source for GPT-6 Luna (opens in a new tab) | |
| 28 | 156.6Confidence interval 154.7–159.2 | Aug 12, 2026Model released | 41active users | Source for Grok 4.6 (opens in a new tab) | |
| 29 | 156.5Confidence interval 154.2–159.3 | Jul 9, 2026Model released | 158active users | Source for GPT-5.6 Luna (opens in a new tab) | |
| 30 | 156.3Confidence interval 153.9–160.0 | Dec 11, 2025Model released | No Tokscale usage in the last 30 days | Source for GPT-5.2 Pro (opens in a new tab) | |
| 31 | 156.0Confidence interval 154.2–158.3 | Aug 14, 2026Model released | 139active users | Source for GLM-5.3 (opens in a new tab) | |
| 32 | 155.9Confidence interval 153.7–158.6 | Sep 9, 2026Model released | 225active users | Source for DeepSeek V4.1 Flash (opens in a new tab) | |
| 33 | 155.7Confidence interval 153.8–158.1 | Feb 5, 2026Model released | 52active users | Source for Claude Opus 4.6 (opens in a new tab) | |
| 34 | 155.6Confidence interval 153.7–158.0 | Aug 13, 2026Model released | 6active users | Source for DeepSeek V4 Pro 0813 (opens in a new tab) | |
| 35 | Gemini 3.1 Pro PreviewListed as Gemini 3.1 Pro | 155.3Confidence interval 153.3–158.1 | Feb 19, 2026Model released | 5active users | Source for Gemini 3.1 Pro Preview (opens in a new tab) |
| 36 | 155.1Confidence interval 152.7–158.0 | Oct 15, 2026Model released | 2active users | Source for Muse Spark Open (opens in a new tab) | |
| 37 | Qwen3.8 Max 0902Listed as Qwen3.8 Max (0902) | 155.1Confidence interval 153.0–157.2 | Sep 1, 2026Model released | 1active users | Source for Qwen3.8 Max 0902 (opens in a new tab) |
| 38 | 155.1Confidence interval 152.8–157.6 | Aug 5, 2026Model released | No Tokscale usage in the last 30 days | Source for Muse Spark 1.2 (opens in a new tab) | |
| 39 | 154.9Confidence interval 152.9–157.4 | May 19, 2026Model released | 3active users | Source for Gemini 3.5 Flash (opens in a new tab) | |
| 40 | 154.8Confidence interval 152.1–156.9 | Jul 31, 2026Model released | 18active users | Source for DeepSeek V4 Flash 0731 (opens in a new tab) | |
| 41 | 154.5Confidence interval 152.8–156.6 | Jul 21, 2026Model released | 3active users | Source for Gemini 3.6 Flash (opens in a new tab) | |
| 42 | 154.4Confidence interval 152.6–157.1 | Sep 21, 2026Model released | 131active users | Source for Grok 4.7 (opens in a new tab) | |
| 43 | 154.1Confidence interval 152.4–156.2 | Jul 8, 2026Model released | 9active users | Source for Grok 4.5 (opens in a new tab) | |
| 44 | 154.1Confidence interval 151.7–156.2 | Jul 9, 2026Model released | No Tokscale usage in the last 30 days | Source for Muse Spark 1.1 (opens in a new tab) | |
| 45 | Qwen3.7 MaxListed as Qwen3.7-Max | 154.0Confidence interval 152.3–156.5 | May 19, 2026Model released | 3active users | Source for Qwen3.7 Max (opens in a new tab) |
| 46 | 153.7Confidence interval 152.1–155.4 | Dec 11, 2025Model released | No Tokscale usage in the last 30 days | Source for GPT-5.2 (opens in a new tab) | |
| 47 | GLM-5.3 FlashListed as GLM-5.3-Flash | 153.1Confidence interval 150.8–155.5 | Aug 20, 2026Model released | 170active users | Source for GLM-5.3 Flash (opens in a new tab) |
| 48 | Gemini 3 Pro PreviewListed as Gemini 3 Pro | 153.1Confidence interval 151.1–155.4 | Nov 18, 2025Model released | 1active users | Source for Gemini 3 Pro Preview (opens in a new tab) |
| 49 | 152.8Confidence interval 150.2–155.1 | Feb 17, 2026Model released | 55active users | Source for Claude Sonnet 4.6 (opens in a new tab) | |
| 50 | Grok 4.20 0309 ReasoningListed as Grok 4.20 | 152.1Confidence interval 149.6–154.5 | Feb 17, 2026Model released | 1active users | Source for Grok 4.20 0309 Reasoning (opens in a new tab) |
Showing 50 of 269 models
Score by release date
Each dot is one model's best result on Epoch Capabilities Index, placed at the model's release date. The dashed line is the best score among models released up to that date.
- Used on Tokscale in the last 30 days
- No Tokscale usage in the last 30 days
- Best score so far
Claude Opus 5.5: 168.8 · released Sep 22, 2026 · 709 active users on Tokscale in the last 30 days
160 models, released Feb 26, 2024 to Oct 15, 2026. Highest: Claude Opus 5.5, 168.8, released Sep 22, 2026. Use the left and right arrow keys to read each model in release order.80100120140160180
Show data
| Model | Released | Score | Tokscale users, 30 days |
|---|---|---|---|
| Muse Spark Open | Oct 15, 2026 | 155.1 | 2 |
| Muse Wonder 1.0 | Oct 12, 2026 | 161.8 | 17 |
| Claude Haiku 5.5 | Oct 7, 2026 | 159.3 | 477 |
| Mistral Large 4 | Oct 6, 2026 | 150.0 | 4 |
| Gemini 4 Argon | Sep 30, 2026 | 166.4 | 82 |
| GPT-6.1 Sol | Sep 29, 2026 | 168.4 | 660 |
| Claude Sonnet 5.5 | Sep 28, 2026 | 166.2 | 555 |
| GPT-6 Luna | Sep 22, 2026 | 156.8 | 433 |
| GPT-6 Sol | Sep 22, 2026 | 163.7 | 246 |
| Claude Opus 5.5 | Sep 22, 2026 | 168.8 | 709 |
| Grok 4.7 | Sep 21, 2026 | 154.4 | 131 |
| DeepSeek V4.1 Flash | Sep 9, 2026 | 155.9 | 225 |
| GPT-6 Astra | Sep 3, 2026 | 168.8 | 563 |
| Muse Spark 1.3 | Sep 2, 2026 | 157.2 | 21 |
| Gemini 3.8 Flash | Sep 2, 2026 | 157.4 | 66 |
| Qwen3.8 Max 0902 | Sep 1, 2026 | 155.1 | 1 |
| Claude Fable 5.1 | Sep 1, 2026 | 166.5 | 361 |
| GLM-5.3 Flash | Aug 20, 2026 | 153.1 | 170 |
| Qwen3.8 27B | Aug 14, 2026 | 151.7 | 6 |
| GLM-5.3 | Aug 14, 2026 | 156.0 | 139 |
| DeepSeek V4 Pro 0813 | Aug 13, 2026 | 155.6 | 6 |
| Gemini 3.7 Flash | Aug 13, 2026 | 158.0 | 4 |
| Grok 4.6 | Aug 12, 2026 | 156.6 | 41 |
| Muse Spark 1.2 | Aug 5, 2026 | 155.1 | No Tokscale usage in the last 30 days |
| Qwen3.8 Max | Aug 2, 2026 | 157.4 | 16 |
| DeepSeek V4 Flash 0731 | Jul 31, 2026 | 154.8 | 18 |
| Qwen3.7 Flash | Jul 27, 2026 | 144.1 | No Tokscale usage in the last 30 days |
| Claude Opus 5 | Jul 24, 2026 | 163.9 | 163 |
| Gemini 3.5 Flash Lite | Jul 21, 2026 | 144.9 | 13 |
| Gemini 3.6 Flash | Jul 21, 2026 | 154.5 | 3 |
| Kimi K3 | Jul 16, 2026 | 157.9 | 56 |
| Inkling | Jul 15, 2026 | 148.4 | No Tokscale usage in the last 30 days |
| Inkling Small | Jul 15, 2026 | 150.3 | No Tokscale usage in the last 30 days |
| Muse Spark 1.1 | Jul 9, 2026 | 154.1 | No Tokscale usage in the last 30 days |
| GPT-5.6 Luna | Jul 9, 2026 | 156.5 | 158 |
| GPT-5.6 Terra | Jul 9, 2026 | 160.1 | 120 |
| GPT-5.6 Sol | Jul 9, 2026 | 162.8 | 217 |
| Grok 4.5 | Jul 8, 2026 | 154.1 | 9 |
| Claude Sonnet 5 | Jun 30, 2026 | 156.9 | 138 |
| GLM-5.2 | Jun 16, 2026 | 151.9 | 13 |
| Kimi K2.7 Code | Jun 12, 2026 | 150.1 | 17 |
| Claude Fable 5 | Jun 9, 2026 | 163.4 | 41 |
| Qwen3.7 Plus | Jun 2, 2026 | 147.0 | 3 |
| MiniMax M3 | Jun 1, 2026 | 146.9 | 26 |
| Claude Opus 4.8 | May 28, 2026 | 159.0 | 222 |
| Qwen3.7 Max | May 19, 2026 | 154.0 | 3 |
| Gemini 3.5 Flash | May 19, 2026 | 154.9 | 3 |
| GPT-5.5 Instant | May 5, 2026 | 142.4 | No Tokscale usage in the last 30 days |
| Mistral Medium 2604 | Apr 28, 2026 | 141.4 | No Tokscale usage in the last 30 days |
| Qwen3.6 Flash | Apr 27, 2026 | 142.9 | No Tokscale usage in the last 30 days |
| DeepSeek V4 Flash | Apr 24, 2026 | 146.1 | 51 |
| DeepSeek V4 Pro | Apr 24, 2026 | 149.1 | 51 |
| GPT-5.5 | Apr 23, 2026 | 159.7 | 30 |
| GPT-5.5 Pro | Apr 23, 2026 | 162.9 | 1 |
| Qwen3.6 27B | Apr 22, 2026 | 146.2 | No Tokscale usage in the last 30 days |
| Qwen3.6 Max Preview | Apr 20, 2026 | 149.4 | No Tokscale usage in the last 30 days |
| Kimi K2.6 | Apr 20, 2026 | 150.9 | 4 |
| Grok 4.3 | Apr 17, 2026 | 148.8 | No Tokscale usage in the last 30 days |
| Claude Opus 4.7 | Apr 16, 2026 | 156.8 | 27 |
| Qwen3.6 35B A3B | Apr 14, 2026 | 143.7 | 1 |
| GLM-5.1 | Apr 7, 2026 | 149.8 | 5 |
| Gemma 4 26B A4B IT | Apr 2, 2026 | 141.9 | 1 |
| Gemma 4 31B IT | Apr 2, 2026 | 143.0 | 3 |
| Qwen3.6 Plus | Mar 31, 2026 | 147.3 | 1 |
| MiniMax M2.7 | Mar 18, 2026 | 145.8 | 6 |
| GPT-5.4 Nano | Mar 17, 2026 | 146.0 | 4 |
| GPT-5.4 Mini | Mar 17, 2026 | 148.7 | 1 |
| GPT-5.4 | Mar 5, 2026 | 157.3 | 1 |
| GPT-5.4 Pro | Mar 5, 2026 | 159.6 | No Tokscale usage in the last 30 days |
| Gemini 3.1 Flash Lite | Mar 3, 2026 | 144.3 | 5 |
| Qwen3.5 Flash | Feb 25, 2026 | 143.5 | No Tokscale usage in the last 30 days |
| Qwen3.5 9B | Feb 24, 2026 | 139.3 | 2 |
| Qwen3.5 35B A3B | Feb 24, 2026 | 142.5 | No Tokscale usage in the last 30 days |
| Gemini 3.1 Pro Preview | Feb 19, 2026 | 155.3 | 5 |
| Grok 4.20 0309 Reasoning | Feb 17, 2026 | 152.1 | 1 |
| Claude Sonnet 4.6 | Feb 17, 2026 | 152.8 | 55 |
| Qwen3.5 Plus | Feb 16, 2026 | 146.2 | No Tokscale usage in the last 30 days |
| Qwen3.5 397B A17B | Feb 13, 2026 | 146.8 | No Tokscale usage in the last 30 days |
| MiniMax M2.5 | Feb 12, 2026 | 146.6 | 3 |
| GLM-5 | Feb 11, 2026 | 145.8 | 2 |
| Claude Opus 4.6 | Feb 5, 2026 | 155.7 | 52 |
| GPT-5.3 Codex | Feb 5, 2026 | 157.1 | 5 |
| Kimi K2.5 | Jan 27, 2026 | 148.1 | No Tokscale usage in the last 30 days |
| GLM-4.7 | Dec 22, 2025 | 143.6 | 4 |
| Gemini 3 Flash Preview | Dec 17, 2025 | 151.9 | 4 |
| GPT-5.2 | Dec 11, 2025 | 153.7 | No Tokscale usage in the last 30 days |
| GPT-5.2 Pro | Dec 11, 2025 | 156.3 | No Tokscale usage in the last 30 days |
| DeepSeek Chat | Dec 1, 2025 | 146.1 | 2 |
| Claude Opus 4.5 | Nov 24, 2025 | 150.1 | 4 |
| Gemini 3 Pro Preview | Nov 18, 2025 | 153.1 | 1 |
| GPT-5.1 | Nov 13, 2025 | 149.8 | No Tokscale usage in the last 30 days |
| Kimi K2 Thinking | Nov 6, 2025 | 145.8 | No Tokscale usage in the last 30 days |
| Claude Haiku 4.5 | Oct 15, 2025 | 142.2 | 102 |
| GPT-5 Pro | Oct 7, 2025 | 150.6 | No Tokscale usage in the last 30 days |
| Claude Sonnet 4.5 | Sep 29, 2025 | 146.7 | 23 |
| Qwen3 Max | Sep 24, 2025 | 142.3 | No Tokscale usage in the last 30 days |
| DeepSeek V3.1 | Aug 21, 2025 | 139.8 | No Tokscale usage in the last 30 days |
| GPT-5 Nano | Aug 7, 2025 | 139.2 | No Tokscale usage in the last 30 days |
| GPT-5 Mini | Aug 7, 2025 | 145.3 | 1 |
| GPT-5 2025 08 07 | Aug 7, 2025 | 150.0 | No Tokscale usage in the last 30 days |
| GPT OSS 20B | Aug 5, 2025 | 137.7 | 2 |
| GPT OSS 120B | Aug 5, 2025 | 139.8 | 6 |
| Claude Opus 4.1 | Aug 5, 2025 | 144.1 | No Tokscale usage in the last 30 days |
| Qwen3 235B A22B Instruct 2507 | Jul 25, 2025 | 138.8 | No Tokscale usage in the last 30 days |
| Kimi K2 Instruct | Jul 12, 2025 | 140.0 | No Tokscale usage in the last 30 days |
| Grok 4 0709 | Jul 9, 2025 | 146.2 | No Tokscale usage in the last 30 days |
| Grok 3 Mini | Jun 24, 2025 | 140.2 | No Tokscale usage in the last 30 days |
| Gemini 2.5 Flash | Jun 17, 2025 | 140.6 | 12 |
| Magistral Small 2506 | Jun 10, 2025 | 133.1 | No Tokscale usage in the last 30 days |
| o3-pro | Jun 10, 2025 | 147.3 | No Tokscale usage in the last 30 days |
| Gemini 2.5 Pro | Jun 5, 2025 | 145.1 | 1 |
| DeepSeek R1 0528 | May 28, 2025 | 141.1 | No Tokscale usage in the last 30 days |
| Claude Sonnet 4 | May 22, 2025 | 141.6 | No Tokscale usage in the last 30 days |
| Claude Opus 4 | May 22, 2025 | 142.5 | No Tokscale usage in the last 30 days |
| Gemini 2.5 Flash Preview 05 20 | May 20, 2025 | 141.4 | No Tokscale usage in the last 30 days |
| Mistral Medium 2505 | May 7, 2025 | 134.0 | No Tokscale usage in the last 30 days |
| Gemini 2.5 Pro Preview 05 06 | May 6, 2025 | 142.2 | No Tokscale usage in the last 30 days |
| Qwen3 30B A3B | Apr 29, 2025 | 136.1 | No Tokscale usage in the last 30 days |
| Qwen3 14B | Apr 29, 2025 | 138.2 | No Tokscale usage in the last 30 days |
| Qwen3 32B | Apr 29, 2025 | 138.4 | No Tokscale usage in the last 30 days |
| Qwen3 8B | Apr 28, 2025 | 136.1 | 1 |
| Qwen3 235B A22B | Apr 28, 2025 | 139.2 | No Tokscale usage in the last 30 days |
| o4-mini | Apr 16, 2025 | 145.5 | No Tokscale usage in the last 30 days |
| o3 | Apr 16, 2025 | 146.7 | No Tokscale usage in the last 30 days |
| GPT-4.1 Nano | Apr 14, 2025 | 129.5 | No Tokscale usage in the last 30 days |
| GPT-4.1 Mini 2025 04 14 | Apr 14, 2025 | 134.9 | 5 |
| GPT-4.1 2025 04 14 | Apr 14, 2025 | 136.7 | 5 |
| Grok 3 Beta | Apr 9, 2025 | 138.2 | No Tokscale usage in the last 30 days |
| Llama 4 Maverick 17B 128e Instruct | Apr 6, 2025 | 132.2 | No Tokscale usage in the last 30 days |
| Gemini 2.5 Pro Exp 03 25 | Mar 31, 2025 | 144.0 | No Tokscale usage in the last 30 days |
| DeepSeek V3 0324 | Mar 24, 2025 | 135.8 | No Tokscale usage in the last 30 days |
| Gemma 3 4B IT | Mar 12, 2025 | 115.9 | No Tokscale usage in the last 30 days |
| Gemma 3 12B IT | Mar 12, 2025 | 123.5 | No Tokscale usage in the last 30 days |
| Gemma 3 27B IT | Mar 12, 2025 | 130.0 | No Tokscale usage in the last 30 days |
| QwQ 32B | Mar 5, 2025 | 137.5 | No Tokscale usage in the last 30 days |
| Claude 3.7 Sonnet | Feb 24, 2025 | 141.0 | No Tokscale usage in the last 30 days |
| o3-mini | Jan 31, 2025 | 140.2 | No Tokscale usage in the last 30 days |
| Qwen Max | Jan 25, 2025 | 132.5 | No Tokscale usage in the last 30 days |
| DeepSeek R1 | Jan 20, 2025 | 138.8 | No Tokscale usage in the last 30 days |
| DeepSeek V3 | Dec 26, 2024 | 132.3 | No Tokscale usage in the last 30 days |
| o1 | Dec 17, 2024 | 141.7 | No Tokscale usage in the last 30 days |
| Llama 3.3 70B Instruct | Dec 6, 2024 | 127.3 | No Tokscale usage in the last 30 days |
| GPT-4o 2024 11 20 | Nov 20, 2024 | 128.7 | 1 |
| Mistral Large 2411 | Nov 18, 2024 | 128.6 | No Tokscale usage in the last 30 days |
| Claude 3.5 Haiku | Oct 22, 2024 | 127.1 | No Tokscale usage in the last 30 days |
| Ministral 3B | Oct 16, 2024 | 118.0 | No Tokscale usage in the last 30 days |
| Llama 3.2 90B Vision Instruct | Sep 24, 2024 | 125.5 | No Tokscale usage in the last 30 days |
| Qwen2.5 Coder 0.5B | Sep 18, 2024 | 87.6 | No Tokscale usage in the last 30 days |
| GPT-4o 2024 08 06 | Aug 6, 2024 | 128.8 | No Tokscale usage in the last 30 days |
| Mistral Large 2407 | Jul 24, 2024 | 127.6 | No Tokscale usage in the last 30 days |
| Llama 3.1 8B Instruct | Jul 23, 2024 | 116.5 | No Tokscale usage in the last 30 days |
| Llama 3.1 70B Instruct | Jul 23, 2024 | 125.9 | No Tokscale usage in the last 30 days |
| Mistral Nemo | Jul 18, 2024 | 118.7 | No Tokscale usage in the last 30 days |
| GPT-4o Mini | Jul 18, 2024 | 126.5 | 1 |
| Gemma 2 27B IT | Jun 24, 2024 | 122.1 | No Tokscale usage in the last 30 days |
| Mistral 7B Instruct V0.3 | May 27, 2024 | 108.9 | No Tokscale usage in the last 30 days |
| GPT-4o 2024 05 13 | May 13, 2024 | 129.0 | No Tokscale usage in the last 30 days |
| GPT-4 Turbo | Apr 9, 2024 | 127.3 | No Tokscale usage in the last 30 days |
| Claude 3 Haiku | Mar 7, 2024 | 118.8 | No Tokscale usage in the last 30 days |
| Mistral Large | Feb 26, 2024 | 122.1 | No Tokscale usage in the last 30 days |
Method and matching
How results are run
- One score per model, with the confidence interval Epoch AI publishes. Epoch fits the index jointly across all results, so a model's score can move when new results are added.
- This is Epoch AI's index, not a Tokscale score. Tokscale does not combine benchmarks.
- The date shown is the model's release date as listed by Epoch AI.
Matching to Tokscale models
- 161 of 270 results are matched to 160 Tokscale models: 128 by exact id, 33 by normalized name.
- No match is marked as reviewed by hand yet.
- 109 results from 109 models are not matched. They stay in the table without a Tokscale link.
- Reasoning effort, thinking budget and agent are conditions of a result. They never change which model a result belongs to.
License and attribution
- CC BY 4.0 (opens in a new tab), retrieved Oct 21, 2026.
- Epoch AI, ‘Capabilities & benchmarking’, epoch.ai/benchmarks (CC BY 4.0)
Sources and attribution
Each score keeps its own source, unit and run conditions. Tokscale does not combine benchmarks into one score, and a missing score means the model was not evaluated, not that it scored zero.