RobustBench-TC - Misleading Description (Abbreviated Name): 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 and an abbreviated name on the correct tool; higher is better.

Metric: Tool-call accuracy (%) under the MisDesc-Abbr perturbation (reward component: MisDesc + abbreviated GT name); 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 2031. 18 models tracked. Updated 7 Oct 2026.

Top models

#ModelSettingScoreOverall rank
1Qwen 3 14BNon-reasoning13.9#1269
2Qwen 3.5 9BNon-reasoning13.2#828
3Qwen 3 8BNon-reasoning12.5#1873
4MUA-RL-14B11.8#1658
5Qwen 3 32BNon-reasoning11.1#972
6DeepSeek R1 Distill Qwen 14B11.1#2771
7O4 Mini10.4#368
8MUA-RL-32B10.4
9MUA-RL-8B10.4#1640
10LoopTool-8B9.7#1473
11LoopTool-32B9
12ToolRL-Llama3.2-3B8.3
13Llama 3.2 3B Instruct7.6#7241
14Qwen 2.5 1.5B Instruct7.6#7470
15ToolRL-Qwen2.5-1.5B7.6#2299

No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=robustbench-tc-misleading-description-abbreviated-name · How It Works · Data refreshed daily, snapshot 2026-11-21.