RobustBench-TC - Misleading Description (Neutral Suffix): leaderboard
RobustBench-TC (2026) tests tool-calling models under deployment noise grounded in real tool-calling failures, using 199 single-turn samples drawn from BFCL, API-Bank, RoTBench, ToolAlpaca and ToolEyes and scored by deterministic rule-based parsers. This board is tool-call accuracy in percent under a misleading description on the correct tool and a neutral suffix on the distractor; higher is better.
Metric: Tool-call accuracy (%) under the MisDesc-N perturbation (reward component: MisDesc + neutral suffix on distractor (_1)); RobustBench-TC samples drawn from BFCL V3 single-turn, API-Bank, RoTBench, ToolAlpaca and ToolEyes, each scored by its source benchmark's strict scorer after a format-tolerant tool-call parse, temperature 0 (DeepSeek-R1-Distill at 0.6), Qwen3-family models with thinking disabled; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 18 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Setting | Score | Overall rank |
|---|---|---|---|---|
| 1 | Qwen 3 32B | Non-reasoning | 63.7 | #983 |
| 2 | O4 Mini | 54.9 | #372 | |
| 3 | LoopTool-8B | 48 | #1497 | |
| 4 | Qwen 3.5 9B | Non-reasoning | 47.1 | #838 |
| 5 | DeepSeek R1 Distill Qwen 14B | 46.1 | #2739 | |
| 6 | MUA-RL-8B | 45.1 | #1656 | |
| 7 | Qwen 3 8B | Non-reasoning | 44.1 | #1890 |
| 8 | MUA-RL-14B | 44.1 | #1674 | |
| 9 | LoopTool-32B | 42.2 | ||
| 10 | MUA-RL-32B | 42.2 | ||
| 11 | Qwen 3 14B | Non-reasoning | 40.2 | #1284 |
| 12 | ToolRL-Qwen2.5-1.5B | 40.2 | #2330 | |
| 13 | ToolRL-Qwen2.5-3B | 39.2 | ||
| 14 | TL-CodeLLaMA-2 | 35.3 | ||
| 15 | Qwen 2.5 3B Instruct | 32.4 | #5410 |
No result here: #2 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=robustbench-tc-misleading-description-neutral-suffix · How It Works · Data refreshed daily, snapshot 2026-11-22.