Geological Reasoning on Geo-Eval
7.35Concept ScoreGPT-5.4 (Frontier)
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| GPT-5.4 (Frontier)Size=-2026.05 | 7.35 | 7.1 | 7 | 7.15 | — | |
| DeepSeek-V3.2-InstructSize=-2026.05 | 6.8 | 6.75 | 6.67 | 6.74 | — | |
| Qwen3-32B-geoSize=32B2026.05 | 6.78 | 6.79 | 6.9 | 6.82 | 1.82 | |
| Gemma-3-27B-geoSize=27B2026.05 | 6.7 | 6.6 | 6.47 | 6.59 | 1.43 | |
| GPT-4oSize=-2026.05 | 6.1 | 5.9 | 5.8 | 5.93 | — | |
| Qwen3-8B-geoSize=8B2026.05 | 6.1 | 6.27 | 6.44 | 6.27 | 1.64 | |
| Gemma-3-27B-ITSize=27B2026.05 | 5.3 | 5.1 | 5.08 | 5.16 | — | |
| Qwen3-32BSize=32B2026.05 | 5.2 | 4.9 | 4.9 | 5 | — | |
| Qwen3-8BSize=8B2026.05 | 4.8 | 4.68 | 4.41 | 4.63 | — | |
| GLM-4-9BSize=9B2026.05 | 4.7 | 4.5 | 4.51 | 4.57 | — | |
| Llama-3.1-70B-InstructSize=70B2026.05 | 4.3 | 4.1 | 3.96 | 4.12 | — |