Opus 5.5 shows 2.3x the research taste of top human experts at ~1/30th the cost
SaxenaNayan · x · 2026-10-07
Periodic Labs reports Opus 5.5 has 2.3x the "experimental research taste" of its best human experts: on a typical task it matches the expert's best score in 17 GPU hours instead of 40. Cost gap is starker — an Opus 5.5 attempt costs $280 (API + GPU) vs $9,123 for an expert attempt, roughly 30x cheaper. Frontier models' experimental research taste has doubled every 3 months since December 2025.
Related event: AI's Ability to Predict Experiments Is Surging, Tests Show(2 posts)→
More from Models
- Mistral's Le Chonk tops blind human code review among open models, second only to Opus 5 — qtnx_ · 2026-10-07
- DeepSeek V4.1 Flash hits 72.9% on ARC-AGI-2 at $0.13/task, costing 250% more — teortaxesTex · 2026-10-07
- Mistral claims Large 4 is one of the world's strongest AI models for cybersecurity — scaling01 · 2026-10-07
- Mistral Large 4 solves 18 of 19 CTF challenges in official speedrun with tool calls — MistralAI · 2026-10-07
- Community poll tiers AI labs: Anthropic and OpenAI frontier, Mistral and Amazon judged 3 generations behind — NathanpmYoung · 2026-10-07
- User: Claude Opus Hit Usage Limit Just 2 Hours Into an Overnight Plotting Run — Sauers_ · 2026-10-07