{"type":"post","id":"kimi-k3-1-pro-leads-terminal-bench-4-a1aa9b","slug":"kimi-k3-1-pro-leads-terminal-bench-4-a1aa9b","title":"Kimi K3.1 Pro leads Terminal-Bench 4.1 at 68%, ahead of Claude Sonnet 5.5 at 66%.","description":"Terminal-Bench 4.1 hardens task verifiers against reward hacking and regrades saved Terminal-Bench 4.0 artifacts without additional inference.","retrieval_nugget":"Terminal-Bench 4.1 hardens task verifiers against reward hacking and regrades saved Terminal-Bench 4.0 artifacts without additional inference. The team says only about 1.16% of 11,550 trials changed score, and all updated agent scores remained within their original 95% confidence intervals. The release matters because benchmark rankings depend on the integrity and fairness of their graders.","published_at":"2026-11-14","updated_at":"2026-11-14","record_date":"2026-11-14","date_kind":"published_at","topics":["claude"],"entities":["Terminal-Bench","Harbor","RewardKit","Epoch AI","Tokenless","Preference Model"],"related_patterns":["ai-engineering-evaluations-are-splitting-into-task-specific"],"novelty_label":"new_fact","what_changed":"Kimi K3.1 Pro now leads Terminal-Bench 4.1 with 68%, ahead of Claude Sonnet 5.5 at 66%.","prior_refs":[],"source_url":"https://www.tbench.ai/news/terminal-bench-4-1","source_urls":["https://www.tbench.ai/news/terminal-bench-4-1"],"schema_version":"newruntime-agent-readable-v0.2","stable_id":"post:kimi-k3-1-pro-leads-terminal-bench-4-a1aa9b","status":"published","visuals":[],"editor":{"role":"editor","name":"Andrey Reshetnikov","profile":"https://newruntime.com/owner-profile.md","profile_json":"https://newruntime.com/owner-profile.json","same_as":["https://www.linkedin.com/in/reshetnikov1/","https://t.me/lapetuse"]},"routes":{"html":"https://newruntime.com/posts/kimi-k3-1-pro-leads-terminal-bench-4-a1aa9b/","markdown":"https://newruntime.com/posts/kimi-k3-1-pro-leads-terminal-bench-4-a1aa9b.md","json":"https://newruntime.com/posts/kimi-k3-1-pro-leads-terminal-bench-4-a1aa9b.json"},"markdown_url":"https://newruntime.com/posts/kimi-k3-1-pro-leads-terminal-bench-4-a1aa9b.md","next_reads":[{"type":"topic","path":"/topics/claude/","url":"https://newruntime.com/topics/claude/","title":"Claude","reason":"Explore the claude topic hub.","media_type":"text/html","markdown_url":"https://newruntime.com/topics/claude.md","json_url":"https://newruntime.com/topics/claude.json"},{"type":"collection","path":"/posts/","url":"https://newruntime.com/posts/","title":"Field Notes","reason":"Open the latest editorial analysis.","media_type":"text/html","json_url":"https://newruntime.com/posts.json"},{"type":"collection","path":"/sources/","url":"https://newruntime.com/sources/","title":"Sources","reason":"See the original sources this archive cites.","media_type":"text/html"}]}