Azalia Mirhoseini(@Azaliamirh):Check out TRACE, a new self-improvement approach where the agent identifies the missing capabilities behind its own failures and trains itself to address them. TRACE-trained Qwen3.6-27B reaches 73.2% on SWE-bench Verified, outperforming much larger models like Codex 5.2 and GLM 5, while beating GRPO and GEPA with <1/4 the training rollouts. By contrasting successful and failed trajectories, TRACE identifies its own weaknesses (such as bug localization or retrieval of the correct doc) and creates new synthetic environments to fix them. The result is a transferable and sample-efficient synthetic env / data generation + fine-tuning pipeline for agentic tasks. Great work led by @TarunSures41845 and @hangoo

2026.07.09 15:46

Check out TRACE, a new self-improvement approach where the agent identifies the missing capabilities behind its own failures and trains itself to address them. TRACE-trained Qwen3.6-27B reaches 73.2% on SWE-bench Verified, outperforming much larger models like Codex 5.2 and GLM 5, while beating GRPO and GEPA with <1/4 the training rollouts. By contrasting successful and failed trajectories, TRACE identifies its own weaknesses (such as bug localization or retrieval of the correct doc) and creates new synthetic environments to fix them. The result is a transferable and sample-efficient synthetic env / data generation + fine-tuning pipeline for agentic tasks. Great work led by @TarunSures41845 and @hangoo_kang!

显示更多

Hangoo Kang@hangoo_kang

2026.07.07 16:39

“TRACE: Capability-Targeted Agentic Training” got Spotlight @ ICML AIWILD 🎉 Beats direct RL, GEPA, & synthetic-agent data on SWE-Bench Verified and τ²-Bench. TRACE-Qwen3.6-27B tops GPT-5.2-Codex, GLM 5, & Claude 4.5 Sonnet on SWE-Bench. Co-led with @TarunSures41845. Thanks to @JonSaadFalcon and our advisor @Azaliamirh. Details below 👇

显示更多