# Vals Terminal-Bench Science 0.1 (Terminal-Bench Science (Vals))

> Vals AI’s fixed-harness, single-trial run of 70 expert-authored scientific research workflows.

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

- Category: [Agentic](/agentic)
- Last updated: September 22, 2026

## About Terminal-Bench Science (Vals)

- Year: 2026
- Tasks: 70 scientific research workflows across five domains
- Format: Strict pass@1 accuracy
- Difficulty: Expert scientific research workflows
- Paper: [Vals Terminal-Bench Science](https://www.vals.ai/benchmarks/terminal-bench-science)

Vals runs all 70 version-0.1 tasks through Terminus 2 and reports strict pass@1 by scientific domain. The fixed harness and single trial differ from the official native-agent, three-trial leaderboard; the Vals rows remain display-only and separate.

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

## Leaderboard (29 models)

| Rank | Model | Configuration | Creator | Score |
|------|-------|---------------|---------|-------|
| 1 | [GPT-6 Astra](/models/gpt-6-astra) | max reasoning | OpenAI | 65.71% |
| 2 | [Claude Opus 5.5](/models/claude-opus-5-5) | — | Anthropic | 48.57% |
| 3 | [Claude Fable 5.1](/models/claude-fable-5-1) | — | Anthropic | 34.29% |
| 4 | [Claude Opus 5](/models/claude-opus-5) | — | Anthropic | 22.86% |
| 5 | [Muse Spark 1.3 Max](https://www.vals.ai/models/meta_muse_spark_1_3_max) | max reasoning | Meta | 14.29% |
| 6 | [Claude Fable 5](/models/claude-fable) | — | Anthropic | 12.86% |
| 7 | [GPT-5.6 Sol](/models/gpt-5-6-sol) | max reasoning | OpenAI | 12.86% |
| 8 | [Grok 4.7](/models/grok-4-7) | xhigh reasoning | xAI | 11.43% |
| 9 | [Claude Opus 4.8](/models/claude-opus-4-8) | — | Anthropic | 10.00% |
| 10 | [GPT-5.6 Terra](/models/gpt-5-6-terra) | max reasoning | OpenAI | 10.00% |
| 11 | [Gemini 3.8 Flash](/models/gemini-3-8-flash) | high reasoning | Google | 8.57% |
| 12 | [Muse Spark 1.3](/models/muse-spark-1-3) | xhigh reasoning | Meta | 8.57% |
| 13 | [Gemini 3.7 Flash](/models/gemini-3-7-flash) | high reasoning | Google | 5.71% |
| 14 | [GPT-5.6 Luna](/models/gpt-5-6-luna) | max reasoning | OpenAI | 5.71% |
| 15 | [Gemini 3.5 Flash](/models/gemini-3-5-flash) | high reasoning | Google | 4.29% |
| 16 | [Grok 4.6](/models/grok-4-6) | high reasoning | xAI | 4.29% |
| 17 | [DeepSeek V4.1 Flash](/models/deepseek-v4-1-flash) | high reasoning | DeepSeek | 4.29% |
| 18 | [GLM-5.3](/models/glm-5-3) | max reasoning | Z.AI | 4.29% |
| 19 | [Claude Sonnet 5](/models/claude-sonnet-5) | — | Anthropic | 2.86% |
| 20 | [MiMo-V2.6-Pro](/models/mimo-v2-6-pro) | — | Xiaomi | 2.86% |
| 21 | [Kimi K3](/models/kimi-k3) | max reasoning | Moonshot AI | 2.86% |
| 22 | [Gemini 3.6 Flash](/models/gemini-3-6-flash) | high reasoning | Google | 1.43% |
| 23 | [Qwen3.8 Max](/models/qwen3-8-max) | — | Alibaba | 1.43% |
| 24 | [Qwen3.8-27B](/models/qwen3-8-27b) | xhigh reasoning | Alibaba | 1.43% |
| 25 | [Hy4 preview](/models/hy4-preview) | — | Tencent | 1.43% |
| 26 | [Mercury 2.5](/models/mercury-2-5) | high reasoning | Inception | 0.00% |
| 27 | [DeepSeek V4 Flash 0731](/models/deepseek-v4-flash-0731) | high reasoning | DeepSeek | 0.00% |
| 28 | [DeepSeek V4 Pro 0813](/models/deepseek-v4-pro-0813) | max reasoning | DeepSeek | 0.00% |
| 29 | [GLM-5.3-Flash](/models/glm-5-3-flash) | max reasoning | Z.AI | 0.00% |

## FAQ

### What does Terminal-Bench Science (Vals) measure?

Vals AI’s fixed-harness, single-trial run of 70 expert-authored scientific research workflows.

### Which model leads the published Terminal-Bench Science (Vals) snapshot?

GPT-6 Astra currently leads the published Terminal-Bench Science (Vals) snapshot with a score of 65.71%.

### How many models are evaluated on Terminal-Bench Science (Vals)?

The September 22, 2026 contains 29 AI models.

### Does Terminal-Bench Science (Vals) affect BenchLM's overall score?

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