Benchmarks / Agentic
GDP.xlsx
Uses Surge AI's first-party spreadsheet-task leaderboard score. Score: score. Collected by Jake Boggs' benchmark dashboard.
GDP.xlsx
Higher is better · best effort per model
full table below
| # | Model | Effort | Score |
|---|---|---|---|
| 1 | Gemini 4 | high | 38.3% |
| 2 | Opus 5.5 | max | 30.3% |
| 3 | Sonnet 5.5 | max | 29.1% |
| 4 | Fable 5.1 | max | 23.1% |
| 5 | GPT-6 Astra | max | 22.9% |
| 6 | Muse Spark 1.3 | max | 22.6% |
| 7 | Grok 4.7 | xhigh | 22.3% |
| 8 | GPT-6.1 Sol | max | 22.0% |
| 9 | GLM 5.3 | max | 21.4% |
| 10 | Qwen 3.8 Max | xhigh | 20.3% |
| 11 | Gemini 3.8 Flash | high | 18.3% |
| 12 | GPT-6 Sol | max | 17.1% |
| 13 | Sonnet 5 | max | 16.9% |
| 14 | GPT-6 Luna | max | 14.9% |
| 15 | Kimi K3 | max | 13.4% |
| 16 | Gemini 3.1 Pro | high | 10.3% |
| 17 | DeepSeek V4 Pro | max | 6.9% |
| 18 | Nemotron 3 Ultra 550B A55B | max | 4.9% |
| 19 | Mistral Large 3 | none | 0.0% |