Skip to main content
Evaluation sources/Terminal-Bench Hard
ObservingCoding

Terminal-Bench Hard

Artificial Analysis · Public model evaluation results from Terminal-Bench Hard.

Official evaluation
Source description

This source measures concrete delivery performance in Coding; raw scores are shown separately from the normalized capability score.

Limitation: Different versions, reasoning settings, and test sets may change scores; read them with the data date.

How it is used

Participating sources affect capability scores; reference, cross-reference, and observing sources are for comparison and verification only.

Model results from this source

Step 3.7 Flash35.635.62026-06-15
Step 3.5 Flash 260332.632.62026-06-02
Step 3.5 Flash27.327.32026-06-02
Gemini 2.5 Pro26.526.52026-06-02
GLM-4.62525.02026-05-22
GLM-4.52222.02026-03-11
GLM-4.5-Air20.520.52026-05-30
Qwen3 Max20.520.52026-05-30
Mistral Small 417.417.42026-06-01
Mistral Large 315.915.92026-06-04
Qwen3-Coder 30B-A3B Instruct15.215.22026-06-02
Gemini 2.5 Flash13.613.62026-06-02
DeepSeek-R16.16.12026-03-11
GLM-4.5V5.35.32026-04-29
Gemini 2.5 Flash-Lite4.54.52026-03-11
Llama-3.3-70B-Instruct33.02026-03-11

Frequently asked questions