ToolSandbox: leaderboard
ToolSandbox: Evaluates tool calling, API use, function selection, structured arguments, and multi-step tool workflows.
Metric: Avg Score (imported). Source: github.com. Source: benchmarklist.com. Saturation forecast: Around October 2026. 17 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Nemotron 3.5 Lightning | 75.13 | #467 |
| 2 | Qwen 3.6 35B A3B | 75.02 | #226 |
| 3 | GPT-4o (2024-05-13) | 73 | #325 |
| 4 | Gemma 4 26B A4B | 72.09 | #358 |
| 5 | Claude 3 Opus (20240229) | 69.2 | #412 |
| 6 | GPT-3.5 Turbo (0125) | 65.6 | #984 |
| 7 | GPT-4 Preview (0125) | 64.3 | #486 |
| 8 | Claude 3 Sonnet (20240229) | 63.8 | #677 |
| 9 | Gemini 1.5 Pro (001) | 60.4 | #444 |
| 10 | Claude 3 Haiku (20240307) | 54.9 | #771 |
| 11 | Gemini 1.0 Pro | 38.1 | #933 |
| 12 | Hermes-2-Pro-Mistral-7B | 31.4 | #1230 |
| 13 | Mistral 7B Instruct (v0.3) | 29.8 | #1274 |
| 14 | c4ai-command-r-v01 | 26.2 | #989 |
| 15 | c4ai-command-r-plus | 24.7 | #724 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=toolsandbox · How It Works · Data refreshed daily, snapshot 2026-10-20.