Environment Alignment on LifeSim
95.9Environment AlignmentDeepSeek-V3.2
Evaluation Results
| Method | Links | |
|---|---|---|
| DeepSeek-V3.22026.03 | 95.9 | |
| GPT-oss-120BParameters=120B2026.03 | 95.9 | |
| DeepSeek-V3.2 ThinkingThinking=true2026.03 | 95.8 | |
| Qwen3-it 32BParameters=32B, Type=Instruction-tuned2026.03 | 95.3 | |
| GPT-52026.03 | 95.1 | |
| GPT-4o2026.03 | 95.1 | |
| GPT-oss-20BParameters=20B2026.03 | 95.1 | |
| Claude-Sonnet-4.52026.03 | 94.9 | |
| Gemma3-it 27BParameters=27B, Type=Instruction-tuned2026.03 | 94.5 | |
| Qwen3-it 14BParameters=14B, Type=Instruction-tuned2026.03 | 94.1 | |
| Gemma3-it 12BParameters=12B, Type=Instruction-tuned2026.03 | 93.8 | |
| Llama3.1-70B-itParameters=70B, Type=Instruction-tuned2026.03 | 93.2 | |
| Qwen3-it 8BParameters=8B, Type=Instruction-tuned2026.03 | 92.6 | |
| Llama3.1-8B-itParameters=8B, Type=Instruction-tuned2026.03 | 91.2 |