Best AI Agents for Knowledge Work. Finance, legal, tax and spreadsheets.
Professional work is judged differently from code. This list averages a model's position on finance, legal, tax and Excel modeling benchmarks, and it is refreshed every week.
Knowledge work ranking
Ranked by average position across four professional-work benchmarks.
As of November 3, 2026, Gemini 4 Carbon holds first place on the knowledge work board, with an average rank of 1.75 across 4 boards, ahead of Gemini 4 Argon (4.25).
Finance, legal, tax and spreadsheet work. A model needs a rank on at least 3 of these 4 boards to be listed.
| Rank | Model | Average rank | Vals Finance Index | Vals Legal Index | Tax Agent Bench | Excel Modeling | Agreement | Week |
|---|---|---|---|---|---|---|---|---|
| 1 |  Google · Proprietary | 1.75 | 1 | 3 | 1 | 2 | Boards agree | – |
| 2 | Google · Proprietary | 4.25 | 3 | 2 | 6 | 6 | Mostly agree | – |
| 3 |  Meta · Proprietary | 4.50 | 4 | 1 | 8 | 5 | Boards disagree | – |
| 4 |  Anthropic · Proprietary | 4.50 | 2 | 8 | 4 | 4 | Mostly agree | – |
| 5 |  Alibaba · Proprietary | 6.25 | 6 | 6 | 2 | 11 | Boards disagree | – |
| 6 | Meta · Proprietary | 8.50 | 7 | 4 | 13 | 10 | Boards disagree | – |
| 7 | Anthropic · Proprietary | 10.00 | 5 | 22 | 10 | 3 | Boards disagree | – |
| 8 |  DeepSeek · Open weights | 10.75 | 8 | 10 | 7 | 18 | Boards disagree | – |
| 9 |  Z.ai · Open weights | 11.00 | 9 | 5 | 11 | 19 | Boards disagree | – |
| 10 |  xAI · Proprietary | 11.00 | 11 | 7 | 14 | 12 | Boards disagree | – |
| 11 |  Moonshot AI · Open weights | 12.75 | 12 | 11 | 15 | 13 | Mostly agree | – |
| 12 | Google · Proprietary | 13.75 | 10 | 16 | 21 | 8 | Boards disagree | – |
| 13 |  Tencent · Open weights | 16.25 | 14 | 12 | 18 | 21 | Boards disagree | – |
| 14 |  Xiaomi · Open weights | 16.50 | 13 | 9 | 19 | 25 | Boards disagree | – |
| 15 | Anthropic · Proprietary | 19.50 | 15 | 14 | 22 | 27 | Boards disagree | – |
| 16 | Z.ai · Open weights | 20.75 | 21 | 13 | 23 | 26 | Boards disagree | – |
| 17 | Xiaomi · Open weights | 22.00 | 16 | 18 | 32 | 22 | Boards disagree | – |
| 18 |  MiniMax · Open weights | 22.00 | 20 | 24 | 20 | 24 | Mostly agree | – |
| 19 |  StepFun · Proprietary | 22.50 | 24 | 15 | 34 | 17 | Boards disagree | – |
| 20 |  Mistral · Proprietary | 22.50 | 17 | 23 | 27 | 23 | Boards disagree | – |
| 21 | Mistral | 23.00 | 18 | 26 | 28 | 20 | Boards disagree | – |
| 22 | Moonshot AI · Proprietary | 24.50 | 19 | 25 | 39 | 15 | Boards disagree | – |
| 23 |  OpenAI · Proprietary | 25.00 | 23 | 21 | 47 | 9 | Boards disagree | – |
| 24 | OpenAI · Proprietary | 25.00 | 22 | 19 | 45 | 14 | Boards disagree | – |
| 25 | DeepSeek · Open weights | 27.00 | 28 | 17 | 35 | 28 | Boards disagree | – |
| 26 | Meta | 28.25 | 27 | 20 | 36 | 30 | Boards disagree | – |
| 27 | OpenAI · Proprietary | 28.75 | 25 | 30 | 53 | 7 | Boards disagree | – |
| 28 | OpenAI · Proprietary | 30.00 | 26 | 28 | 50 | 16 | Boards disagree | – |
| 29 |  Inception | 35.00 | 30 | 27 | 51 | 32 | Boards disagree | – |
| 30 | MiniMax · Open weights | 36.33 | 29 | – | 49 | 31 | Boards disagree | – |
| 31 |  Reflection · Open weights | 38.25 | 31 | 29 | 64 | 29 | Boards disagree | – |
| 32 |  Thinking Machines · Open weights | 38.75 | 34 | 31 | 55 | 35 | Boards disagree | – |
| 33 | Thinking Machines · Open weights | 39.00 | 33 | 33 | 56 | 34 | Boards disagree | – |
| 34 |  NVIDIA · Open weights | 39.25 | 32 | 32 | 60 | 33 | Boards disagree | – |
The boards behind this ranking
- Vals Finance Index by Vals AIAnalyst research, financial modeling and corporate tax research. 34 models. Updated Oct 30.
- Vals Legal Index by Vals AILegal research and drafting across practice areas. 33 models. Updated Oct 30.
- Tax Agent Bench by Vals AIResearch-grade US corporate tax questions. 77 models. Updated Oct 30.
- Excel Modeling by Vals AIExcel-based financial modeling tasks. 35 models. Updated Oct 29.
43 more models appear on too few of these boards to average fairly, so they are not ranked here.
Sources: Arena leaderboard dataset (CC BY 4.0; we keep each model's best configuration and renumber), Terminal-Bench, Vals AI benchmarks. Re-read every week. How we calculate it. eSEOspace does not run these benchmarks and has not tested these models for this ranking. The order comes from the formula, and no one can pay for a position. eSEOspace is a member of the Claude Partner Network, and Claude is one of the models ranked here.
How to read it
How to read this ranking
Read the average rank first: lower is better. Then look at the Agreement column. A model the boards agree on is a safer bet than one that is first on one board and twentieth on another.
Select a model's name to see its score on each board and the harness used. Use the tabs to switch from the overall list to coding, knowledge work or reliability, and the filters to narrow by lab or to open-weight models.
Other views: Overall. Coding. Reliability. Methodology.
Limits
What this ranking can't tell you
A high average tells you a model did well on these tests. It does not prove it will handle your work, your data or your tools. Run your own trial on real tasks before you commit.
Boards do not cover every model. A model missing from a tough board can look better than one that was tested and scored lower, which is why we show how many boards each model appears on and whether they agree.
Benchmarks can also be gamed. Researchers have shown that agents can sometimes score well by exploiting a test rather than solving it, so treat a small gap between two models as a tie.
A strong rank on legal or tax tasks is not legal or tax advice. A qualified professional still needs to review anything these models produce.
FAQ
Common questions
Which AI is best for finance and legal work?
Can an AI agent replace an analyst or a paralegal?
Next steps
Go deeper
AI Agent vs Chatbot vs Automation: What Is Different
Read the guide →Model vs Agent vs Harness: Why the Same AI Scores Differently
Read the guide →AI Agent Cost Per Task: Why Token Price Misleads
Read the guide →AI Agent Frameworks Compared: Claude SDK, OpenAI, LangGraph
Read the guide →Best AI Agents for Accounting and Bookkeeping: 9 Options
Read the guide →Best AI Customer Support Agents: 9 Options Compared
Read the guide →All AI agent guides
See all guides →
Project Managers who will work with you on your project!
Irina S.hello@eseospace.com
Benjamin G.ben@eseospace.com
David G.david@eseospace.comWork with us
Talk to the team behind your project
Tell us what you need and we’ll match you with the right people.
Book a Discovery Call
Jeanette M.jeanette@eseospace.com
Mark S.mark@eseospace.com
Maryna P.maryna@eseospace.com
Wondering what this costs? Every service has published pricing, and you don't need a discovery call to see it.
View pricingPicking an AI agent for your business?
A benchmark rank tells you which models are strong. It does not tell you which one fits your data, your tools and your risk. Tell us what you want the agent to do and we will give you a straight answer.
Talk to our team
