ALTK-Evolve tops customer-service benchmark via learned memory

Hugging Face introduced ALTK-Evolve, which uses GPT-4.1 to distill short behavioral rules from an agent's training-task successes and failures and append them to the system prompt, without updating model weights. On Sierra's τ²-bench retail customer-service tasks, a basic ReAct agent reached 67.5% Pass^4, beating every retail entry on the leaderboard snapshot.

• Agent Consistency: Hugging Face's New Diagnostic2026-09-15
• Train Microduck to walk in Isaac Lab 3.0 and deploy to real robot2026-10-21
• IBM releases Granite Speech 5.0 speech model family2026-10-21
• Hugging Face launches SimReady Central with 2,000+ simulation-ready 3D assets2026-10-21
• XHToken releases Spark-X2.6-4B on-device model2026-10-21