Terminal-Bench 2.1 (Claude Code)

Terminal-Bench 2.1 evaluated using Claude Code 2.1.126, parser=json, temp=1.0, top_p=1.0, max_new_tokens=131072, avg of 5 runs.

Category

Agentic Coding

Max Score

100.0

Score Type

percent

Active

Yes

Model Scores

Scores for Terminal-Bench 2.1 (Claude Code)

Model Score Date Verified
Ornith-1.0-9B
43.75%
25.06.2026 Verified
Qwen3.5-9B
18.90
25.06.2026 Unverified
Qwen3.5-35B-A3B
40.32%
25.06.2026 Unverified
Ornith-1.0-35B
88.51%
25.06.2026 Verified
Qwen3.6-35B-A3B
61.09%
25.06.2026 Unverified
Qwen3.5-397B-A17B
59.88%
25.06.2026 Unverified
Ornith-1.5-35B-A3B
100.00%
19.08.2026 Verified
Ornith-1.0-35B-A3B
88.51%
19.08.2026 Unverified