RobustBench-TC - Timeout: 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 first tool call that returns a timeout; higher is better.

Metric: Tool-call accuracy (%) under the Timeout perturbation (transition component: first tool call returns “Tool execution timed out”); 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

#ModelSettingScoreOverall rank
1Qwen 3 14BNon-reasoning72.9#1269
2Qwen 3.5 9BNon-reasoning71.4#828
3LoopTool-8B69.8#1473
4Qwen 3 8BNon-reasoning66.8#1873
5MUA-RL-32B64.3
6O4 Mini62.3#368
7MUA-RL-8B58.8#1640
8MUA-RL-14B57.3#1658
9Qwen 2.5 3B Instruct54.3#5392
10LoopTool-32B54.3
11ToolRL-Qwen2.5-3B47.7
12Qwen 3 32BNon-reasoning47.2#972
13ToolRL-Llama3.2-3B43.7
14Qwen 2.5 1.5B Instruct43.2#7470
15DeepSeek R1 Distill Qwen 14B42.7#2771

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-timeout · How It Works · Data refreshed daily, snapshot 2026-11-21.