DAYJOB: Tax
Professional Work · 2026-10-27
Surge AI DAYJOB: Tax evaluates professional judgment on 50 expert-designed tax assignments spanning 1,789 workspace files across eight practice areas, from individual, small-business and partnership returns to estates, international tax and disputes. An attempt passes when it satisfies at least 95% of the task's rubric criteria, and the leaderboard reports pass@1 over repeated attempts; Surge's case studies describe five attempts per task. The Tax page and launch post name no harness or grader; the launch post places Tax in the DAYJOB family, whose published cross-domain harness is the OpenHands SDK in Harbor with no browsing or package downloads.
Top models (higher is better)
| Model | Score |
|---|---|
| Claude Sonnet 5.5 | 10.8 |
| Claude Fable 5.1 | 9.6 |
| Claude Opus 5.5 | 9.2 |
| Muse Wonder | 6.0 |
| Grok 4.7 | 5.0 |
| GPT-6 Astra | 4.8 |
| GPT-6.1 Sol | 4.0 |
| Muse Spark 1.3 | 3.6 |
| Gemini 4 Argon | 3.2 |
| DeepSeek V4.1 Flash 1017 | 1.2 |
| Gemini 3.8 Flash | 0.8 |
| GLM-5.3 | 0.4 |
| Kimi K3 | 0.0 |
| Mistral Large 4 Preview | 0.0 |