RobustBench-TC - Schema Drift: leaderboard
Metric: Tool-call accuracy (%) under the SchemaD perturbation (transition component: first call returns “parameter X no longer valid”); RobustBench-TC samples drawn from BFCL V3 single-turn, API-Bank, RoTBench, ToolAlpaca and ToolEyes, each scored by its source benchmark's strict scorer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 2032. 18 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | O4 Mini | 52.3 |
| 2 | Qwen 3 8B (Non-reasoning) | 40.7 |
| 3 | DeepSeek R1 Distill Qwen 14B | 40.7 |
| 4 | Qwen 3 14B (Non-reasoning) | 40.7 |
| 5 | Qwen 2.5 1.5B Instruct | 35.7 |
| 6 | Qwen 2.5 3B Instruct | 35.7 |
| 7 | Qwen 3 32B (Non-reasoning) | 33.7 |
| 8 | Llama 3.2 3B Instruct | 28.6 |
| 9 | Qwen 3.5 9B (Non-reasoning) | 17.6 |
Interactive version: theaggregate.ai/benchmark?slug=robustbench-tc-schema-drift · How It Works · Data refreshed daily, snapshot 2026-10-07.