List your company

Software Engineering on Terminal-Bench 2.0

75.1Public ScoreGPT-5.4
51.1857.3963.669.81Jun 11, 2026

Evaluation results

MethodLinks
GPT-5.4Harness=Proprietary2026.0675.1
Claude Opus 4.7Harness=Terminus-22026.0669.4
Gemini 3.1 ProHarness=Terminus-22026.0668.5
GPT-5.2Harness=Codex CLI2026.0662.9
Claude Opus 4.6Harness=Claude Code2026.0658
Qwen3.5-397B-A17BHarness=Terminus-22026.0652.5
Claude Opus 4.6Harness=OpenCode2026.0652.1