Terminal-based task execution on Terminal-Bench 2.0
64.7AccuracyGPT-5.3-Codex
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| GPT-5.3-CodexModel Category=Proprietary, Scaffold=Terminus 22026.04 | 64.7 | — | |
| Claude Opus 4.6Model Category=Proprietary, Scaffold=Terminus 22026.04 | 62.9 | — | |
| Claude Opus 4.5Model Category=Proprietary, Scaffold=Terminus 22026.04 | 57.8 | — | |
| Gemini 3 ProModel Category=Proprietary, Scaffold=Terminus 22026.04 | 56.9 | — | |
| GPT-5.2Model Category=Proprietary, Scaffold=Terminus 22026.04 | 54 | — | |
| GLM 5Model Category=Open-Source, Scaffold=Terminus 22026.04 | 52.4 | — | |
| GPT-5.1Model Category=Proprietary, Scaffold=Terminus 22026.04 | 47.6 | — | |
| Kimi K2.5Model Category=Open-Source, Scaffold=Terminus 22026.04 | 43.2 | — | |
| Claude Sonnet 4.5Model Category=Proprietary, Scaffold=Terminus 22026.04 | 42.8 | — | |
| MiniMax m2.5Model Category=Open-Source, Scaffold=Terminus 22026.04 | 42.2 | — | |
| DeepSeek-V3.2Model Category=Open-Source, Scaffold=Terminus 22026.04 | 39.6 | — | |
| Claude Opus 4.1Model Category=Proprietary, Scaffold=Terminus 22026.04 | 38 | — | |
| GPT-5Model Category=Proprietary, Scaffold=Terminus 22026.04 | 35.2 | — | |
| Qwen3-32B + SSModel Category=Open-Source, Scaffold=Terminus 2, Training Trajectories=SkillSynth2026.04 | 29.6 | — | |
| Claude Haiku 4.5Model Category=Proprietary, Scaffold=Terminus 22026.04 | 28.3 | — | |
| GPT-5-MiniModel Category=Proprietary, Scaffold=Terminus 22026.04 | 24 | — | |
| Qwen 3 Coder 480BModel Category=Open-Source, Scaffold=Terminus 22026.04 | 23.9 | — | |
| Qwen3-14B + SSModel Category=Open-Source, Scaffold=Terminus 2, Training Trajectories=SkillSynth2026.04 | 19.9 | — | |
| Qwen3-8B + SSModel Category=Open-Source, Scaffold=Terminus 2, Training Trajectories=SkillSynth2026.04 | 13.5 | — | |
| AnteModel=Gemini 3 Pro2026.02 | — | 64.7 | |
| Claude CodeModel=Claude Opus 4.5, ADK=Claude2026.02 | — | 52.1 | |
| Codex CLIModel=GPT-5.2 (xhigh), ADK=OpenAI2026.02 | — | 62.9 | |
| OpenHandsModel=Claude Opus 4.5, ADK=OpenHands2026.02 | — | 51.9 | |
| SageAgentModel=Gemini 3 Pro, ADK=OpenSage2026.02 | — | 65.2 |