🧰 SWE-bench Verified

The items that name it 19

• [AINews] OpenAI’s Patch the Planet program is 1 of 2 parties to discover PHP2Shell vulnerabilities in core PHP affecting 70% of the web, using GPT-5.6 Sol; coordinated response over 2 months
  article2026-10-22
• [AINews] OpenAI’s Patch the Planet program is 1 of 2 parties to discover PHP2Shell vulnerabilities in core PHP affecting 70% of the web, using GPT-5.6 Sol; coordinated response over 2 months
  article2026-10-22
• VibeTraining-Bench: Data-Trainer-Harness Co-Design for Iterative System Improvement
  paper2026-10-19
• Master the captain (@mollicksoif9) on X
  thread2026-10-17
• LFM2.5-24B-A2B: Built for fast on-device agents
  article2026-10-14
• [AINews] AI Engineer NYC: Gemini 4 Argon, Vals AI Benchmarks
  article2026-10-14
• Show HN: Parsimony – an open-source benchmark for how much code AI models add
  tool2026-10-13
• Run fast, private agents on-device with LFM2.5-24B-A2B
  article2026-10-13
• Anthropic says AI distillation attack defenses cut Claude extraction by 90% to 99%
  article2026-10-13
• Strengthening our defenses against distillation attacks
  article2026-10-13
• Strengthening our defenses against distillation attacks
  article2026-10-12
• Before They Can Solve: Predicting Post-Training Coding-Agent Performance from Base Models
  paper2026-10-07
• Aleph Alpha Kolibri: How the sovereign German LLM works
  article2026-10-03
• TokenCast: Forecasting Token Consumption During LLM Agent Execution
  paper2026-09-28
• Shockingly Simple Self-retrospection Improves Agentic Models Without RL
  paper2026-09-28
• trailhq/Graft
  repo2026-07-03
• Aleph Alpha Kolibri-1 (78B MoE, Apache 2.0)
  repo
• Same Model, Different Harness: Different Coding-Agent Results
  paper
• Same Model, Different Harness: Different Coding-Agent Results
  paper