Terminal-Bench
Updated
Cite
Fill in the access date before citing.
Plain text
Benchmark Registry. Terminal-Bench 2.1 — mini-swe-agent benchmark results. Benchmark identifier: terminal-bench/2-1-mini-swe-agent. https://benchmarkregistry.org/benchmarks/terminal-bench/2-1-mini-swe-agent. Accessed [YYYY-MM-DD]. Data license: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Attribution: Benchmark Registry, a project of Densa Labs — https://benchmarkregistry.org/ (CC BY 4.0). Include the cited record URL and identify any changes.
BibTeX
@misc{br_terminal_bench_2_1_mini_swe_agent,
title = {Terminal-Bench 2.1 — mini-swe-agent benchmark results},
publisher = {Benchmark Registry},
url = {https://benchmarkregistry.org/benchmarks/terminal-bench/2-1-mini-swe-agent},
urldate = {YYYY-MM-DD},
note = {Benchmark identifier: terminal-bench/2-1-mini-swe-agent; data license: CC BY 4.0; access date to be filled by the reader}
}Results
3 results
| Provider | Model | Score | Source | Registry No. |
|---|---|---|---|---|
| Meta AI (originally Facebook AI Research) | Muse Spark 1.1 (xhigh) | 80.0% | dev.meta.ai for Muse Spark 1.1 (xhigh) on Terminal-Bench 2.1 — mini-swe-agent (opens in a new tab)Terminal-Bench AccuracyReported Model pageMetaSelf-reportedRecord for Muse Spark 1.1 on Terminal-Bench 2.1 — mini-swe-agent | 80002 |
| Anthropic | Claude Sonnet 5 (xhigh) | 80.4% | www-cdn.anthropic.com for Claude Sonnet 5 (xhigh) on Terminal-Bench 2.1 — mini-swe-agent (opens in a new tab)Terminal-Bench AccuracyReported System cardAnthropicSelf-reportedRecord for Claude Sonnet 5 on Terminal-Bench 2.1 — mini-swe-agent | 20013 |
| Anthropic | Claude Fable 5 (high) | 84.3% | www-cdn.anthropic.com for Claude Fable 5 (high) on Terminal-Bench 2.1 — mini-swe-agent (opens in a new tab)Terminal-Bench AccuracyReported System cardAnthropicSelf-reportedRecord for Claude Fable 5 on Terminal-Bench 2.1 — mini-swe-agent | 20012 |