Atlas
Benchmarks
326 sources · 1,256 models × 2,956 benchmarks · 154,484 scoresScience's Last Exam: Protocol Planning
Whether an agent understands how operations alter experimental states and can plan reasonable next steps, scored as mean task reward × 100 over 40 tasks.
Models
6
Top score
58.0
Leader
GLM-5.3 Flash
Updated
Nov 4, 2026
Resolution
—
Saturation
—
No signal or difficulty estimate. Too few models have been measured on this benchmark to estimate its difficulty.
| # | Model | Setting | Score | Source |
|---|---|---|---|---|
| 1 | GLM-5.3 Flash | high · Science's Last Exam; Claude Code with web search and fetch off; DiscoveryRSI for Hypothesis Discovery | 58.0 | |
| 2 | Kimi K3.1 Pro | not reported · Science's Last Exam; Claude Code with web search and fetch off; DiscoveryRSI for Hypothesis Discovery | 58.0 | |
| 3 | Claude Opus 5 | high · Science's Last Exam; Claude Code with web search and fetch off; DiscoveryRSI for Hypothesis Discovery | 55.0 | |
| 4 | GPT-6.1 Sol | medium · Science's Last Exam; Codex with web search and fetch off; DiscoveryRSI for Hypothesis Discovery | 51.0 | |
| 5 | MiMo-V2.6-Pro | not reported · Science's Last Exam; Claude Code with web search and fetch off; DiscoveryRSI for Hypothesis Discovery | 49.0 | |
| 6 | MiniMax M3 | high · Science's Last Exam; Claude Code with web search and fetch off; DiscoveryRSI for Hypothesis Discovery | 43.0 |
