Excel Modeling Benchmark Dataroom
Professional Work · 2026-07-01
This row reports the category score for dataroom-style tasks in Excel Modeling Benchmark.
Top models (higher is better)
| Model | Score |
|---|---|
| Claude Opus 5 | 87.7 |
| Opus 4.8 | 85.9 |
| Claude Fable 5 | 85.4 |
| GPT-5.6 Sol | 85.2 |
| Kimi K3 | 85.1 |
| Sonnet 5 | 84.9 |
| GLM-5.2 | 83.4 |
| GPT-5.6 Luna | 83.0 |
| GPT-5.6 Terra | 81.6 |
| GPT-5.5 | 81.2 |
| Gemini 3.5 Flash | 80.9 |
| Qwen3.7-Max | 80.7 |
| Gemini 3.1 Pro Preview | 79.2 |
| Muse Spark 1.1 | 78.6 |
| Sonnet 4.6 | 78.5 |
| MiMo-V2.5-Pro | 78.0 |
| Kimi K2.6 | 75.1 |
| Gemini 3.6 Flash | 73.0 |
| GPT-5.4 Nano | 72.0 |
| Qwen3.7-Plus | 71.7 |
| Grok 4.5 | 65.0 |
| DeepSeek-V4-Pro | 62.6 |
| Gemini 3.5 Flash-Lite | 57.7 |
| MiniMax M3 | 52.7 |
| Inkling | 51.9 |
| GPT-5.4 Mini | 48.3 |
| Inkling-Small | 37.3 |
| Grok 4.3 | 29.2 |
| Gemini 3.1 Flash-Lite Preview | 15.2 |