RTI Study - Qwen3 Generation References - Tau2 Airline: leaderboard
Untrained reference-model rows on Navigation, three tau2 domains and Claw-Eval Hard20. Pass@1 only; generation-matched comparison cohorts remain separate. Fine-tuned GRPO/GSPO/AEPO/OPD/LUFFY/RTI checkpoints are not attributed to their base models. Source inference conditions are retained in the paper.
Metric: Pass@1 (%). Source: arxiv.org. Saturation forecast: Estimated already saturated. 4 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Sonnet 4 | 54 | #198 |
| 2 | GPT-4.1 | 53 | #243 |
| 3 | Qwen 3 235B A22B | 50.5 | #306 |
| 4 | Qwen 3 8B | 31 | #669 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #7 Gemini 4 Argon.
Interactive version: theaggregate.ai/benchmark?slug=rti-study-qwen3-generation-references-tau2-airline · How It Works · Data refreshed daily, snapshot 2026-10-16.