Long-Context Understanding on Long-Context Understanding
66.8ScoreGPT-5
Evaluation results
| Method | Links | ||
|---|---|---|---|
| GPT-5Generation Mode=LLM-in-Sandbox2026.01 | 66.8 | 0.5 | |
| GPT-5Generation Mode=Standard LLM2026.01 | 66.3 | — | |
| DeepSeek-V3.2-ThinkingGeneration Mode=LLM-in-Sandbox2026.01 | 63.8 | 3 | |
| Claude-Sonnet-4.5-ThinkGeneration Mode=LLM-in-Sandbox2026.01 | 61.8 | 1.3 | |
| Kimi-K2-ThinkingGeneration Mode=LLM-in-Sandbox2026.01 | 61.8 | 1.5 | |
| DeepSeek-V3.2-ThinkingGeneration Mode=Standard LLM2026.01 | 60.8 | — | |
| Claude-Sonnet-4.5-ThinkGeneration Mode=Standard LLM2026.01 | 60.5 | — | |
| Kimi-K2-ThinkingGeneration Mode=Standard LLM2026.01 | 60.3 | — | |
| MiniMax-M2Generation Mode=LLM-in-Sandbox2026.01 | 58.5 | 6.2 | |
| MiniMax-M2Generation Mode=Standard LLM2026.01 | 52.3 | — | |
| Qwen3-Coder-30B-A3BGeneration Mode=Standard LLM2026.01 | 24.6 | — | |
| Qwen3-Coder-30B-A3BGeneration Mode=LLM-in-Sandbox2026.01 | 23.9 | -0.7 | |
| Qwen3-4B-Instruct-2507Generation Mode=LLM-in-Sandbox2026.01 | 10.5 | 1.9 | |
| Qwen3-4B-Instruct-2507Generation Mode=Standard LLM2026.01 | 8.6 | — |