Terminal-Bench-Science leaderboard launches: GPT-6 Astra 63%, Opus 5.5 62%, open weights ~10%
CurieuxExplorer · x · 2026-09-25
Artificial Analysis launched its leaderboard for Terminal-Bench-Science 0.1, an agentic benchmark for scientific research work. GPT-6 Astra (max) leads at 63% and Claude Opus 5.5 (xhigh) at 62% — the only two models above 50%; open-weights models remain around 10%.
- Built by Steven Dillmann and researchers at Stanford with the Terminal-Bench team and a global open-source community, announced August 2026
- 70 expert-curated tasks based on real research across five domains: life sciences (19), physical (17), mathematical (17), engineering (9), earth sciences (8)
- Each task drops an agent into a sandbox with data, tools and instructions, mirroring the Terminal-Bench setup
More from Models
- Gemini 4 Carbon matches predecessor's score but explores less, gets distracted — patience_cave · 2026-11-11
- Fable 5.2 takes MazeBench frontier, matches GPT-6 Astra puzzle-solving — patience_cave · 2026-11-11
- MazeBench: Fable 5.2 scores 15%, near-frontier at half price; Gemini 4 Carbon 11% — patience_cave · 2026-11-11
- ERNIE 5.5 Flash Preview enters LMArena Battle Mode for crowdsourced ranking — arena · 2026-11-11
- Indie dev flips from $20 Pro back to $200 Gemini Ultra after two days of deep use — yihui_indie · 2026-11-11
- OpenAI retiring ChatGPT's Standard Voice with 9 original voices on Dec 11, petition launched — ChaosGremlinOG · 2026-11-11