DAREBench - Text - Multi-Step with Tools: leaderboard
Metric: Accuracy (%). Source: arxiv.org. 35 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Opus 4.8 | 65.3 |
| 2 | MiMo-V2.5-Pro | 61.2 |
| 3 | Claude Sonnet 4.6 | 59.9 |
| 4 | GPT-5.5 | 59.8 |
| 5 | GPT-5.4 | 59.6 |
| 6 | MiMo-V2-Pro | 58.4 |
| 7 | Claude Opus 4.6 | 58.3 |
| 8 | MiMo-V2.5 | 57.5 |
| 9 | Gemini 3.1 Pro (Preview) | 56.1 |
| 10 | Qwen 3.6 Plus | 54.8 |
| 11 | DeepSeek V4 Pro | 54.3 |
| 12 | Qwen 3.5 Plus | 53.9 |
| 13 | GLM-5 | 53.3 |
| 14 | MiniMax-M2.7 | 53 |
| 15 | Gemini 3.5 Flash | 51.7 |
Interactive version: theaggregate.ai/benchmark?slug=darebench-text-multi-step-with-tools · How It Works · Data refreshed daily, snapshot 2026-09-19.