Benchmarks / Agentic
EnterpriseBench CoreCraft
Uses Surge AI's first-party enterprise-task leaderboard score. Score: score. Collected by Jake Boggs' benchmark dashboard.
EnterpriseBench CoreCraft
Higher is better · best effort per model
full table below
| # | Model | Effort | Score |
|---|---|---|---|
| 1 | Fable 5.1 | max | 77.4% |
| 2 | Opus 5.5 | max | 74.4% |
| 3 | Fable 5 | max | 70.3% |
| 4 | Opus 5 | max | 68.7% |
| 5 | Grok 4.7 | xhigh | 68.2% |
| 6 | Grok 4.6 | xhigh | 65.6% |
| 7 | Muse Spark 1.3 | xhigh | 61.9% |
| 8 | Gemini 3.8 Flash | high | 58.5% |
| 9 | Qwen 3.8 Flash | xhigh | 57.4% |
| 10 | Gemini 3.7 Flash | high | 55.9% |
| 11 | DeepSeek V4 Pro | max | 53.3% |
| 12 | Opus 4.8 | max | 52.3% |
| 13 | GLM 5.3 | max | 51.8% |
| 14 | Sonnet 5 | max | 51.3% |
| 15 | GPT-5.5 | xhigh | 51.3% |
| 16 | Gemini 3.5 Flash | high | 50.8% |
| 17 | GLM 5.3 Flash | max | 46.7% |
| 18 | GPT-5.6 Sol | max | 46.2% |
| 19 | DeepSeek V4 Flash | max | 45.6% |
| 20 | GPT-6 Luna | max | 43.1% |
| 21 | GPT-6 Sol | max | 43.1% |
| 22 | Kimi K3 | max | 43.1% |
| 23 | GPT-5.2 | xhigh | 42.6% |
| 24 | GPT-6 Astra | max | 42.6% |
| 25 | GLM 5.2 | xhigh | 42.1% |
| 26 | DeepSeek V4.1 Flash | max | 41.5% |
| 27 | Hy3 | high | 41.0% |
| 28 | Gemini 3.6 Flash | high | 39.5% |
| 29 | GPT-5.4 | xhigh | 36.4% |
| 30 | Opus 4.7 | max | 35.9% |
| 31 | Opus 4.6 | max | 30.8% |
| 32 | Gemini 3.1 Pro | high | 27.2% |
| 33 | Qwen 3.7 Max | max | 26.2% |
| 34 | Kimi K2.6 | max | 24.6% |
| 35 | DeepSeek V3.2 | high | 24.1% |
| 36 | Grok 4.1 Fast | high | 20.5% |
| 37 | GPT-5.2 Codex | xhigh | 20.1% |
| 38 | Gemini 3 Flash | high | 20.0% |
| 39 | GLM 5 | high | 17.4% |
| 40 | Sonnet 4.6 | max | 16.4% |
| 41 | Gemini 3 Pro | high | 14.4% |
| 42 | Qwen 3.5 Plus | none | 11.3% |
| 43 | Kimi K2.5 | max | 8.7% |
| 44 | Mistral Large 3 | none | 3.6% |
| 45 | Qwen 3 Max | max | 3.6% |