{"slug": "altk-evolve-how-learned-memory-tops-the-t2-bench-leaderboard-on-customer-service", "title": "ALTK-Evolve: How learned memory tops the τ²-bench leaderboard on customer service tasks without a frontier model", "summary": "IBM Research's ALTK-Evolve reached 67.5% Pass^4 on the τ²-bench retail benchmark using GPT-4.1, topping the leaderboard without a frontier model or any weight updates. The method automatically distills a ReAct agent's own failed conversations into system-prompt behavioral rules, letting teams improve policy-bound customer service agents through offline prompt evolution instead of fine-tuning or model swaps. The approach shifts the optimization bottleneck from model capability to system-prompt token length, according to the Hugging Face post.", "body_md": "[Hugging Face](https://huggingface.co/blog/ibm-research/altk-evolve-bc)\n\n### ALTK-Evolve: How learned memory tops the τ²-bench leaderboard on customer service tasks without a frontier model\n\nWhich summary reads better? Pick one — models revealed after.Both summaries are AI-generated.\n\nALTK-Evolve achieves 67.5% Pass^4 (solves tasks correctly four times consecutively) on τ²-bench retail using GPT-4.1 by distilling past attempts into behavioral rules, proving lightweight, in-context learning can match frontier-model performance for policy-bound tasks. This lets teams upgrade existing agents without costly fine-tuning or model swaps by dynamically generating and applying domain-specific guidelines—cutting deployment time and compute costs while maintaining strict compliance.\n\nBy automatically distilling its own conversation failures into system-prompt behavioral rules, a basic ReAct agent using the older GPT-4.1 model achieved 67.5% Pass^4 on the τ²-bench retail benchmark, outperforming all frontier entries without any weight updates. This proves that policy-bound customer service agents can be optimized for reliability and strict guardrails entirely through automated offline prompt evolution rather than costly fine-tuning or upgrading to reasoning models. For teams shipping agents, this enables a self-improving pipeline where models dynamically write their own domain playbooks, shifting the optimization bottleneck from model capability to system-prompt token length.", "url": "https://wpnews.pro/news/altk-evolve-how-learned-memory-tops-the-t2-bench-leaderboard-on-customer-service", "canonical_source": "https://www.snipvote.com/story/cmvhsfpl10004yw9jo9kg650i", "published_at": "2026-10-21 07:40:28.055783+00:00", "updated_at": "2026-10-21 07:40:29.891633+00:00", "lang": "en", "topics": ["ai-agents", "large-language-models", "ai-research", "ai-tools"], "entities": ["IBM Research", "ALTK-Evolve", "GPT-4.1", "τ²-bench", "Hugging Face", "ReAct"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/altk-evolve-how-learned-memory-tops-the-t2-bench-leaderboard-on-customer-service", "markdown": "https://wpnews.pro/news/altk-evolve-how-learned-memory-tops-the-t2-bench-leaderboard-on-customer-service.md", "text": "https://wpnews.pro/news/altk-evolve-how-learned-memory-tops-the-t2-bench-leaderboard-on-customer-service.txt", "jsonld": "https://wpnews.pro/news/altk-evolve-how-learned-memory-tops-the-t2-bench-leaderboard-on-customer-service.jsonld"}}