FrontierRSI benchmark: Claude Opus 5.5 leads at 34.3%, GPT-6 models trail by ~10 points
nathanbenaich · x · 2026-11-19
Anthropic researchers believe Claude delivers them a bigger speedup than OpenAI researchers get from GPT — and the newly released FrontierRSI benchmark results are consistent with that. Leaderboard (avg@2): Claude Opus 5.5 (max) / Claude Code tops at 34.3%, Claude Fable 5.2 follows at 33.0%; Gemini 4 Carbon 29.8% and Gemini 4 Argon 23.6% (both with Antigravity CLI); GPT-6.1 Sol 23.5% and GPT-6 Astra 20.9% (with Codex); Kimi K3.1 Pro 13.6%, Muse Wonder 1.0 12.0%, Grok 4.7 just 8.9%.
More from coding & agent
- MemSkill: Memory Skills Learned with LLaMA Transfer to Qwen Without Retraining — jiqizhixin · 2026-11-19
- Honcho Launches Dream Histories to Inspect How Its Memory System 'Dreams' About Users — Scobleizer · 2026-11-19
- Six-year PKM coach: a markdown dump folder isn't a second brain — build a schema.md for your agent — evielync · 2026-11-19
- RL-tuned Qwen3.8-27B for opencode hits Mac via MLX at 156 tok/s on M5 Max — yb2698 · 2026-11-19
- Elon Musk praises Google's Neon 3D-modeling a photorealistic lab via Blender MCP in Antigravity — elonmusk · 2026-11-19
- Jeremy Howard flags Gemini CLI MCP limits: 10k output cap, wrong cwd — jeremyphoward · 2026-11-19