TaxEval v2
Professional Work · 2025-04-18
Tax-question benchmark from Vals evaluating model answers to realistic tax questions and response-writing tasks.
Top models (higher is better)
| Model | Score |
|---|---|
| Muse Spark 1.1 | 79.7 |
| Muse Spark | 77.7 |
| Sonnet 4.6 | 77.1 |
| Claude Fable 5 | 76.9 |
| GPT-5.6 Luna | 76.2 |
| GPT-5.6 Terra | 76.2 |
| Opus 4.6 | 76.0 |
| Grok 3 | 75.9 |
| GPT-5.2 | 75.8 |
| Kimi K3 | 75.7 |
| Grok 4 Fast | 75.7 |
| Opus 4.8 | 75.6 |
| Sonnet 5 | 75.6 |
| Inkling | 75.3 |
| Qwen3.7-Max | 75.3 |
| Opus 4.7 | 75.3 |
| GPT-5 Mini | 75.2 |
| Claude Opus 5 | 75.1 |
| GPT-4.1 | 75.1 |
| GPT-5.5 | 75.0 |
| Gemini 3.6 Flash | 74.9 |
| GPT-5.1 | 74.9 |
| Opus 4.5 | 74.9 |
| O4 Mini | 74.8 |
| GPT-5.6 Sol | 74.8 |
| Qwen3.6 Plus (2026-04-02) | 74.7 |
| Kimi K2.6 | 74.7 |
| o3 | 74.6 |
| GPT-4o (2024-11-20) | 74.5 |
| Gemini 3.5 Flash | 74.4 |
| O1 | 74.3 |
| Kimi K2.5 | 74.2 |
| Grok 4.20 | 74.1 |
| Claude 3.7 Sonnet | 74.0 |
| GPT-5.4 | 74.0 |
| Qwen3 Max Preview | 74.0 |
| Gemini 3 Flash Preview | 73.9 |
| MiMo-V2.5-Pro | 73.8 |
| Opus 4.1 | 73.7 |
| Qwen3 Max (rolling alias) | 73.5 |