Surge AI benchmark: every Claude model, even Haiku, beats every GPT on staff-level coding
echen · x · 2026-11-14
Surge AI benchmarked top Claude and GPT models on sudo L7, its coding benchmark for staff+ engineer judgment and behavior. Results:
- Claude Sonnet 5.5: 62.2% (top)
- Claude Opus 5.5: 57.9%
- Claude Haiku 5.5: 49.3%
- GPT-6 Astra: 47.7%
- GPT-6.1 Sol: 47.4%
Every Claude model — including the smallest Haiku — outscored every GPT tested, giving the Claude lineup a clear edge on staff-level engineering judgment.
More from Models
- Together AI webinar dissects Ripple-65B coding model: dense vs MoE, ablations, RL runs — togethercompute · 2026-11-14
- Developer seeks free model alternatives after losing free DeepSeek access in Pi editor — nirurin · 2026-11-14
- Skeptic doubts Anthropic achieved large-factor bpb gains at fixed compute — teortaxesTex · 2026-11-14
- AI cracks Gödel's 1,009-page notebooks, uncovering hundreds of unknown theorems in days — stevenstrogatz · 2026-11-14
- NVIDIA lands nvfp4 in large-scale pretraining under codename "yolo bundle" — SonglinYang4 · 2026-11-14
- Delip Rao mocks vendors testing safety in production, then rolling back updates — deliprao · 2026-11-14