Surge AI benchmarks agents on 50 professional tax assignments
Original title: Dayjob: Tax – evaluating AI agents on 50 professional tax assignments
Claude Sonnet 5.5 leads DAYJOB: Tax at a 10.8% pass rate under a 95% rubric threshold.
Surge AI released DAYJOB: Tax, 50 environments across eight tax-practice areas where agents receive client records and must produce returns, workpapers, reviews, or advice. Tasks require spotting consequential issues the client did not ask about. At pass@1 with a 95% rubric threshold, Claude Sonnet 5.5 scored 10.8%, Claude Fable 5.1 9.6%, and Claude Opus 5.5 9.2%.
Unconfirmed: Scores and case examples are Surge AI's; the published write-up is truncated.
- The benchmark contains 1,789 workspace files, with a median of 30 per assignment.
- Coverage includes individual, partnership, international, corporate, and dispute work.
- Sonnet 5.5 and Opus 5.5 sit at the high end of the reported cost frontier.
- In one example, Opus caught a home-sale deadline Astra missed in five attempts.
- Products
- DAYJOB: Tax
- Organizations
- Surge AI
From the original article
6 min read at surgehq.ai
Anna was planning to sell her condo. She had moved out in August 2024 and wanted to know whether closing in October 2026 or January 2027 would be better for taxes. An AI agent reviewed both options and correctly concluded that either date would satisfy the residence requirement for a tax break worth up to $250,000 in excluded gains.
By landonxi · surgehq.ai
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.