Question Answering on SteuerEx 1.0 (n=115)
399Total PointsDeepSeek-R1-671B
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| DeepSeek-R1-671BParameters=671B2026.02 | 399 | — | 0.0003 | |
| SteuerLLMParameters=28B2026.02 | 294 | — | — | |
| Open-SteuerLLM2026.02 | 241 | — | 0.1901 | |
| Gemma-3-27B-itParameters=27B, Instruction-tuned=true2026.02 | 236 | — | 0.0029 | |
| GPT-4o-mini2026.02 | 226 | — | 0.0029 | |
| Mistral-Small-it-2409Version=24092026.02 | 212 | — | 0.0001 | |
| DeepSeek-R1-Distill-Llama-70BParameters=70B, Base Model=Llama2026.02 | 204 | — | 0.0001 | |
| Qwen2.5-72B-itParameters=72B, Instruction-tuned=true2026.02 | 196 | — | 0.0001 | |
| Qwen2.5-32B-itParameters=32B, Instruction-tuned=true2026.02 | 186 | — | 0.0001 | |
| Small-SteuerLLM2026.02 | 171 | — | 0.0001 | |
| Ministral-8B-it-2410Parameters=8B, Version=24102026.02 | 136 | — | 0.0001 | |
| Qwen2.5-14B-itParameters=14B, Instruction-tuned=true2026.02 | 129 | — | 0.0001 | |
| Qwen2.5-7B-itParameters=7B, Instruction-tuned=true2026.02 | 120 | — | 0.0001 | |
| Gemma-3-4B-itParameters=4B, Instruction-tuned=true2026.02 | 113 | — | 0.0001 | |
| Llama-3.2-3B-itParameters=3B, Instruction-tuned=true2026.02 | 90 | — | 0.0001 | |
| Llama-3-8B-itParameters=8B, Instruction-tuned=true2026.02 | 79 | — | 0.0001 | |
| Qwen2.5-3B-itParameters=3B, Instruction-tuned=true2026.02 | 54 | — | 0.0001 |