Issues

• OpenAI shuts down o4-mini on 2026-10-23 (and o3-2025-04-16 on 2026-12-11): default model arguments

  meta-pytorch/KernelAgent#152 · Filter by author otisz22 opened on Oct 4, 2026

• [Question] Project roadmap and future development plans

  meta-pytorch/KernelAgent#149 · Filter by author wan3x opened on Sep 3, 2026

• BeamSearchStrategy: silent worker starvation when num_expanding_parents=0

  meta-pytorch/KernelAgent#145 · Filter by author jayden1711 opened on Jul 20, 2026

• Clarification on reported 1.56x speedup aggregation method

  meta-pytorch/KernelAgent#144 · Filter by author XAheli opened on Jul 3, 2026

• Regarding the choice of large language models?

  meta-pytorch/KernelAgent#122 · Filter by author Arthur-Ling opened on Mar 18, 2026
  3

• Support DeepSeek as a LLM Provider

  meta-pytorch/KernelAgent#67 · Filter by author can-gaa-hou opened on Dec 25, 2025
  1

  

• KernelBench Format Support

  enhancement
  meta-pytorch/KernelAgent#63 · Filter by author sandlbn opened on Dec 15, 2025
  1

• regex sandbox bypass via dynamic import / getattr

  enhancementllm-cheesing
  meta-pytorch/KernelAgent#62 · Filter by author sandlbn opened on Dec 14, 2025
  2

• Wrong solution to L3-31_VisionAttention in the artifact repo

  meta-pytorch/KernelAgent#52 · Filter by author mzweilin opened on Dec 3, 2025
  6