# Terminal-Bench-Science 0.1

> A benchmark of AI agents completing expert-curated research workflows across the life, physical, Earth, mathematical, and engineering sciences.

Canonical page: https://benchlm.ai/benchmarks/terminal-bench-science

- Category: [Agentic](/agentic)
- Last updated: October 15, 2026 snapshot

## About Terminal-Bench-Science 0.1

- Year: 2026
- Tasks: 70 expert-curated scientific research workflows
- Format: Resolution rate across 3 trials per task
- Difficulty: Frontier scientific research workflows
- Paper: [Terminal-Bench-Science 0.1](https://www.terminal-bench-science.ai/)

The first release contains 70 workflows contributed and reviewed by researchers across five scientific domains. Each model-and-agent system runs three independent trials per task and produces concrete artifacts graded by reproducible, task-specific tests. We mirror the official system leaderboard as a standalone display-only benchmark.

Terminal-Bench-Science 0.1 is currently displayed on BenchLM for reference, but it is excluded from the weighted scoring formula.

## Leaderboard (17 models)

| Rank | Model | Configuration | Creator | Score |
|------|-------|---------------|---------|-------|
| 1 | [GPT-6 Astra](/models/gpt-6-astra) | Codex · max reasoning | OpenAI | 68.1% |
| 2 | [Opus 5.5](https://www.terminal-bench-science.ai/) | Claude Code · max reasoning | Anthropic | 63.3% |
| 3 | [Claude Fable 5.1](/models/claude-fable-5-1) | Claude Code · max reasoning | Anthropic | 40.0% |
| 4 | [Claude Opus 5](/models/claude-opus-5) | Claude Code · max reasoning | Anthropic | 30.0% |
| 5 | [GPT-5.6 Sol](/models/gpt-5-6-sol) | Codex · max reasoning | OpenAI | 22.4% |
| 6 | [Claude Fable 5](/models/claude-fable) | Claude Code · max reasoning | Anthropic | 21.4% |
| 7 | [DeepSeek V4.1 Flash 1017](/models/deepseek-v4-1-flash) | Codex · max reasoning | DeepSeek | 15.7% |
| 8 | [Grok 4.7](/models/grok-4-7) | Grok Build · xhigh reasoning | xAI | 14.3% |
| 9 | [Gemini 3.8 Flash](/models/gemini-3-8-flash) | mini-SWE-agent · high reasoning | Google | 12.4% |
| 10 | [Claude Opus 4.8](/models/claude-opus-4-8) | Claude Code · max reasoning | Anthropic | 10.5% |
| 11 | [GPT-5.6 Terra](/models/gpt-5-6-terra) | Codex · max reasoning | OpenAI | 8.6% |
| 12 | [GLM-5.3](/models/glm-5-3) | Claude Code · max reasoning | Z.AI | 8.1% |
| 13 | [Kimi K3](/models/kimi-k3) | Claude Code · max reasoning | Moonshot AI | 7.1% |
| 14 | [Grok 4.6](/models/grok-4-6) | Grok Build · xhigh reasoning | xAI | 7.1% |
| 15 | [Gemini 3.7 Flash](/models/gemini-3-7-flash) | mini-SWE-agent · high reasoning | Google | 5.7% |
| 16 | [DeepSeek V4 Pro 0813](/models/deepseek-v4-pro-0813) | Codex · max reasoning | DeepSeek | 3.8% |
| 17 | [GPT-5.6 Luna](/models/gpt-5-6-luna) | Codex · max reasoning | OpenAI | 3.3% |

## FAQ

### What does Terminal-Bench-Science 0.1 measure?

A benchmark of AI agents completing expert-curated research workflows across the life, physical, Earth, mathematical, and engineering sciences.

### Which model leads the published Terminal-Bench-Science 0.1 snapshot?

GPT-6 Astra currently leads the published Terminal-Bench-Science 0.1 snapshot with a score of 68.1%.

### How many models are evaluated on Terminal-Bench-Science 0.1?

The October 15, 2026 snapshot contains 17 AI models.

### Does Terminal-Bench-Science 0.1 affect BenchLM's overall score?

Not directly. Terminal-Bench-Science 0.1 is still displayed on BenchLM for reference, but it is excluded from the weighted scoring formula.