Agentic on BrowseComp
78.4ScoreKimi K2.5
Evaluation results
| Method | Links | |
|---|---|---|
| Kimi K2.5Agent Architecture=Agent Swarm2026.02 | 78.4 | |
| GLM-5Context Management=w/ Context Manage2026.02 | 75.9 | |
| Kimi K2.5Context Management Protocol=w/ ctx manage2026.02 | 74.9 | |
| Kimi K2.5Context Management=w/ Context Manage2026.02 | 74.9 | |
| Gemini 3 ProContext Management Protocol=w/ ctx manage2026.02 | 67.6 | |
| DeepSeek-V3.2Context Management=w/ Context Manage2026.02 | 67.6 | |
| GLM-4.7Context Management=w/ Context Manage2026.02 | 67.5 | |
| GPT-5.2 (xhigh)Context Management Protocol=standard2026.02 | 65.8 | |
| GPT-5.2 (xhigh)Context Management=w/ Context Manage2026.02 | 65.8 | |
| GLM-5Context Management=None2026.02 | 62 | |
| Kimi K2.5Context Management Protocol=standard2026.02 | 60.6 | |
| Kimi K2.5Context Management=None2026.02 | 60.6 | |
| GPT-5.2 (xhigh)Context Management Protocol=w/ ctx manage2026.02 | 59.2 | |
| Gemini 3 ProContext Management=w/ Context Manage2026.02 | 59.2 | |
| Claude Opus 4.5Context Management Protocol=w/ ctx manage2026.02 | 57.8 | |
| Claude Opus 4.5Context Management=w/ Context Manage2026.02 | 57.8 | |
| GPT-5 HighVariant=High2026.01 | 54.9 | |
| GLM-4.7Context Management=None2026.02 | 52 | |
| DeepSeek-V3.2Context Management Protocol=standard2026.02 | 51.4 | |
| DeepSeek-V3.2 ThinkingThinking Mode=true2026.01 | 51.4 | |
| DeepSeek-V3.2Context Management=None2026.02 | 51.4 | |
| MiMo-V2-FlashVariant=Flash2026.01 | 45.4 | |
| Gemini 3 ProContext Management Protocol=standard2026.02 | 37.8 | |
| Gemini 3 ProContext Management=None2026.02 | 37.8 | |
| Claude Opus 4.5Context Management Protocol=standard2026.02 | 37 | |
| Claude Opus 4.5Context Management=None2026.02 | 37 | |
| Claude Sonnet 4.5Variant=Sonnet 4.52026.01 | 24.1 |