← All stories
● Covered by 1 source · 1 reportMedium impact1 neutral

Surge AI Releases DAYJOB: Tax Benchmark for Evaluating AI Agents on Professional Tax Assignments

🔄 Updated 9d ago
New to BrevFeed? We gather this story from every outlet covering it into one summary — ranked by real-world impact, not just the latest headline — so you never miss what matters. What is BrevFeed? →

Key points

  • Surge AI launched DAYJOB: Tax benchmark.
  • Evaluates AI agents on 50 professional tax assignments.
  • Tests specialized expertise and professional judgment.
  • Claude Sonnet 5.5 leads with 10.8% pass rate.

Introduction of DAYJOB: Tax

Surge AI has released DAYJOB: Tax, the latest addition to its DAYJOB benchmark family. This new benchmark focuses on evaluating AI agents' performance in professional tax assignments. It aims to test AI capabilities beyond simple instruction following, requiring specialized expertise and complex decision-making in tax preparation, compliance, and advisory services.

Benchmark Design and Evaluation

The DAYJOB: Tax benchmark comprises 50 distinct environments covering eight areas of tax practice. AI agents are presented with client requests and supporting records, then tasked with producing deliverables such as tax returns, workpapers, technical reviews, or client advice. The evaluation criteria include tax knowledge, calculations, retrieval, instruction following, and crucially, professional judgment, such as recognizing consequential issues not explicitly requested by the client.

Example Scenario and Performance Metrics

An example scenario involved a client seeking advice on condo sale timing for tax benefits. While one AI agent (Astra) failed to identify a critical deadline that could affect eligibility for a tax exclusion, another (Opus) consistently identified it. Performance is measured by a pass@1 rate at a 95% rubric threshold, assessing calculations, supporting evidence, consequential issues, and the completeness of the deliverable.

Initial Leaderboard Results

Initial results from the DAYJOB: Tax leaderboard show Claude Sonnet 5.5 leading with a 10.8% pass rate. Claude Fable 5.1 follows with 9.6%, and Claude Opus 5.5 achieved a 9.2% pass rate. These figures indicate the current state of AI performance in handling complex professional tax tasks.

✨ This summary was generated by AI from the outlets' reporting listed below. It is not independently verified and may contain errors — check the original sources. How BrevFeed works →

The daily brief

One email each morning: the day's tech stories, clustered across outlets and summarized. No account needed.

One email a day. Unsubscribe in one click, any time.

Today's brief

Spend a few minutes, get the whole day. Every topic's top stories in one hands-free rundown — listen, watch, or read the transcript.

~4 min · 3 stories · Nov 07

▶ Play today's brief Listen on Spotify

New every morning, and the back catalogue is archived by date.

Reporting from

Surge AI introduced DAYJOB: Tax, a new benchmark to evaluate AI agents on 50 professional tax assignments. This benchmark assesses AI's ability to handle complex tax scenarios requiring specialized expertise and professional judgment, extending Surge AI's DAYJOB family of benchmarks.