# Terminal-Bench-Science 0.1 ![Category](https://img.shields.io/badge/Category-Agent-1f6feb) ![Metric](https://img.shields.io/badge/Metric-%25%20Accuracy-2da44e) ![Version](https://img.shields.io/badge/Version-0.1-8957e5) ![Saturation](https://img.shields.io/badge/Saturation-Low-d29922) > **Quick answer:** Terminal-Bench-Science 0.1 is an early-version benchmark that measures how well AI agents perform agentic scientific-research tasks through a terminal interface, reporting accuracy against mean cost per task across multiple effort/reasoning levels. Claude Fable 5.1 scores 52.6% (at "max" effort) as of September 2026, per Anthropic's public reporting. ## At a Glance **What it tests:** An AI agent's ability to complete real scientific-research workflows — using a terminal/shell environment as the interaction surface — rather than isolated science-knowledge Q&A. **Why it matters:** Most science benchmarks test static knowledge recall; Terminal-Bench-Science 0.1 instead measures whether an agent can actually execute a multi-step research task end-to-end, which is closer to how frontier labs expect models to accelerate real scientific work. **Known limitations:** Version 0.1 signals this is an early-stage benchmark — task count and long-term stability are not yet publicly documented. Anthropic reports a standard error of ±3.5–4.5 points per model, and notes its own reproduction of published scores can differ slightly from a separate public leaderboard using a different harness (3 trials/task, Claude Code harness). ## What Terminal-Bench-Science 0.1 Measures Terminal-Bench-Science 0.1 evaluates an AI agent's ability to carry out agentic scientific-research tasks — the kind of multi-step, tool-using workflow a research assistant might perform — through a terminal/shell interface, rather than through static question answering. Results are typically reported as an accuracy-vs-cost curve across several reasoning-effort tiers (low, medium, high, xhigh, max), since higher-effort settings trade increased compute cost for higher accuracy. Anthropic introduced this benchmark data point in its September 2026 Claude Fable 5.1 / Claude Mythos 5.1 announcement, reporting both its own "max effort" scores for Fable 5.1 and Fable 5, and noting that a separate public leaderboard (using 3 trials per task inside a Claude Code harness) reports different absolute numbers for Claude Opus 5 and Claude Fable 5 — within the benchmark's stated noise band, but a reminder that harness choice materially affects Terminal-Bench-Science 0.1 results. ## Benchmark Specifications | Field | Value | |---|---| | Task category | Agent (scientific research) | | Metric | % accuracy (reported vs. mean cost per task, USD) | | Version | 0.1 | | Saturation | Low | | Reported standard error | ±3.5–4.5 pts per model | ## How Terminal-Bench-Science 0.1 Is Scored Agents attempt scientific-research tasks inside a terminal environment and are scored on the percentage completed correctly. Because task difficulty allows models to spend more inference compute for higher accuracy, scores are commonly reported at multiple reasoning-effort levels (low/medium/high/xhigh/max) alongside the mean dollar cost per task at each level — making Terminal-Bench-Science 0.1 as much a cost-efficiency benchmark as a pure accuracy one. ## State-of-the-Art Results | Rank | Model | Score | Source | Date | |---|---|---|---|---| | 1 | [Claude Fable 5.1](/models/anthropic/claude-fable-5-1) | 52.6% (max effort) | [Anthropic: Claude Fable 5.1 and Mythos 5.1](https://www.anthropic.com/claude-fable-and-mythos-5-1) | 2026-09 | | 2 | [Claude Opus 5](/models/anthropic/claude-opus-5) | 29.0% (Anthropic reproduction) / 30.0% (public leaderboard) | [Anthropic: Claude Fable 5.1 and Mythos 5.1](https://www.anthropic.com/claude-fable-and-mythos-5-1) | 2026-09 | | 3 | [Claude Fable 5](/models/anthropic/claude-fable-5) | 24.7% (Anthropic reproduction, max) / 21.4% (public leaderboard) | [Anthropic: Claude Fable 5.1 and Mythos 5.1](https://www.anthropic.com/claude-fable-and-mythos-5-1) | 2026-09 | | 4 | [GPT-5.6 Sol](/models/openai/gpt-5-6-sol) | 22.4% | [Anthropic: Claude Fable 5.1 and Mythos 5.1](https://www.anthropic.com/claude-fable-and-mythos-5-1) | 2026-09 | *Scores sourced from Anthropic's Claude Fable 5.1 / Claude Mythos 5.1 announcement, September 2026. Fable 5.1's score is reported at "max" reasoning effort; lower-effort tiers trade accuracy for lower per-task cost (Fable 5.1: 26.3% at low effort up to 52.6% at max).* ## Terminal-Bench-Science 0.1 on Benchgen > No Benchgen results yet — [be the first to run Terminal-Bench-Science 0.1](/benchmarks/anthropic/terminal-bench-science-01). ## Terminal-Bench-Science 0.1 vs Other Benchmarks | Benchmark | What it tests | Saturation | |---|---|---| | **Terminal-Bench-Science 0.1** | Agentic scientific-research tasks via terminal use | Low | | [Terminal-Bench 4.0](/benchmarks/anthropic/terminal-bench-40) | General agentic terminal/coding tasks | Low | | [SciCode](/benchmarks/scicode) | Scientific code-generation from research problems | Low | | [FrontierScience Research](/benchmarks/frontierscience-research) | Broader frontier scientific-research task suite | Low | ## Run Terminal-Bench-Science 0.1 on Your Model Benchgen lets you track agentic scientific-research performance across reasoning-effort tiers and harnesses, complementing vendor-reported Terminal-Bench-Science 0.1 scores with independently repeatable, cost-aware evaluation. ## Frequently Asked Questions

What is Terminal-Bench-Science 0.1?

Terminal-Bench-Science 0.1 is an early-version benchmark measuring how well AI agents complete agentic scientific-research tasks through a terminal interface, reported as accuracy versus mean cost per task.

What does a good score look like on Terminal-Bench-Science 0.1?

Claude Fable 5.1 reports 52.6% at max reasoning effort as of September 2026 — the highest publicly reported score at benchmark launch, with a stated standard error of ±3.5–4.5 points per model.

Who created Terminal-Bench-Science 0.1?

The benchmark's authorship isn't publicly detailed in Anthropic's launch announcement; Anthropic references a separate public leaderboard (3 trials/task, Claude Code harness) that also reports scores for this benchmark.

Rank	Model	Score
1	gpt-6-astra	64.6
2	claude-opus-5-5	58.7
3	claude-fable-5-1	52.6

Terminal-Bench-Science 0.1 — Results

Terminal-Bench-Science 0.1 — Evaluations