Tool Use on MCPMark
57.5Total Success RateGPT-5.2 High
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| GPT-5.2 High2026.06 | 57.5 | — | — | — | — | — | |
| Seed 2.0Model Category=Frontier Proprietary Models2026.04 | 54.7 | 60 | 39.1 | 53.6 | 40 | 81 | |
| Seed2.0 Pro2026.06 | 54.7 | — | — | — | — | — | |
| Gemini-3-pro High2026.06 | 53.9 | — | — | — | — | — | |
| GPT-5.2 HighModel Category=Frontier Proprietary Models2026.04 | 53.1 | 60 | 47.8 | 42.9 | 40 | 66.7 | |
| GPT-5 Highenvironment=internal2025.12 | 50.9 | — | — | — | — | — | |
| Gemini-3 ProModel Category=Frontier Proprietary Models2026.04 | 50.8 | 56.7 | 45.7 | 43.8 | 40 | 70.2 | |
| Gemini-3.0 Proenvironment=internal2025.12 | 43.1 | — | — | — | — | — | |
| Claude-Opus-4.52026.06 | 42.3 | — | — | — | — | — | |
| CODEDELEGATORBackbone=DeepSeekV3.2, Temperature=0, Metric=pass@1 (%)2026.01 | 38.4 | 49.2 | 38 | 34.8 | 27 | 41.7 | |
| DeepSeek-V3.2environment=internal, thinking mode=true2025.12 | 38 | — | — | — | — | — | |
| DeepSeek-V3.2-685BModel Category=Open-Source Foundation Models, Parameter Scale=685B2026.04 | 36.7 | 36.7 | 20.7 | 45.5 | 17 | 66.6 | |
| Claude-4.5-Sonnetenvironment=internal2025.12 | 33.3 | — | — | — | — | — | |
| Claude Sonnet-4.5Model Category=Frontier Proprietary Models2026.04 | 33.3 | 32.5 | 29.4 | 25 | 27 | 50 | |
| Claude-Sonnet-4.52026.06 | 32.1 | — | — | — | — | — | |
| CodeActBackbone=DeepSeekV3.2, Temperature=0, Metric=pass@1 (%)2026.01 | 26.4 | 39 | 27.9 | 19.4 | 9.7 | 36.2 | |
| ReActBackbone=DeepSeekV3.2, Temperature=0, Metric=pass@1 (%)2026.01 | 25.8 | 35 | 18.5 | 13.4 | 13 | 52.4 | |
| MiniMax M2environment=internal2025.12 | 24.4 | — | — | — | — | — | |
| Kimi-K2environment=internal, thinking mode=true2025.12 | 20.4 | — | — | — | — | — | |
| Agent-World-14BModel Category=Open-Source Environment Scaling Methods, Parameter Scale=14B2026.04 | 13.3 | 16.6 | 4.4 | 3.6 | 4 | 38.1 | |
| Agent-World-8BModel Category=Open-Source Environment Scaling Methods, Parameter Scale=8B2026.04 | 8.9 | 13.3 | 4.4 | 3.6 | 4 | 19.1 | |
| Qwen3-32BModel Category=Open-Source Foundation Models, Parameter Scale=32B2026.04 | 7.5 | 10 | 0 | 3.6 | 0 | 23.8 | |
| Qwen3-235B-A22BModel Category=Open-Source Foundation Models, Parameter Scale=235B2026.04 | 5.8 | 13.3 | 0 | 10.7 | 0 | 4.8 | |
| EnvScaler-8BModel Category=Open-Source Environment Scaling Methods, Parameter Scale=8B2026.04 | 5.6 | 10 | 4.4 | 0 | 4 | 9.5 | |
| AWM-14BModel Category=Open-Source Environment Scaling Methods, Parameter Scale=14B2026.04 | 5.1 | 3.3 | 8.7 | 0 | 4 | 9.5 | |
| GPT-OSS-120BModel Category=Open-Source Foundation Models, Parameter Scale=120B2026.04 | 4.7 | 5.8 | 4.4 | 3.6 | 3 | 7.1 | |
| Qwen3-14BModel Category=Open-Source Foundation Models, Parameter Scale=14B2026.04 | 3.4 | 3.3 | 4.4 | 0 | 0 | 9.5 | |
| Qwen3-8BModel Category=Open-Source Foundation Models, Parameter Scale=8B2026.04 | 2.4 | 3.3 | 0 | 0 | 4 | 4.8 | |
| Simulator-8BModel Category=Open-Source Environment Scaling Methods, Parameter Scale=8B2026.04 | 2.4 | 3.3 | 0 | 0 | 4 | 4.8 | |
| AWM-8BModel Category=Open-Source Environment Scaling Methods, Parameter Scale=8B2026.04 | 2.4 | 3.3 | 0 | 0 | 4 | 4.8 | |
| TOUCAN-7BModel Category=Open-Source Environment Scaling Methods, Parameter Scale=7B2026.04 | 1 | 0 | 0 | 0 | 0 | 4.8 |