Leaderboard
FinDeepResearch
Whether deep-research agents and LLMs can write a rigorous corporate financial analysis report that follows an expert-designed structure. Grading covers each analyst step: recognising data, calculating metrics, abstracting, and interpreting.
Best reported
41.9
FinAgentTest
overall accuracy score (0-100)
23 Mar 2026
FinDeepResearch Paper v2 (2026-01-07); latest board entry 2026-03-23
overall accuracy score (0-100), higher is better. 10 models.
Updated 23 Mar 2026
- FinAgentTest41.9
- Gemini 3 Pro Deep Research38
- OpenAI o3-deep-research37.9
- grok-437.3
- Gemini 2.5 Pro36.2
- OpenAI GPT-533
- Gemini 3 Pro32.7
- Tongyi Deep Research32.1
- claude-sonnet-4.525.8
- Perplexity Sonar Deep Research24.7
FinDeepResearch Paper v2 (2026-01-07); latest board entry 2026-03-23
overall accuracy score (0-100), higher is better. 10 models.
Updated 23 Mar 2026
- 1FinAgentTest41.9
- 2Gemini 3 Pro Deep Research38
- 3OpenAI o3-deep-research37.9
- 4grok-437.3
- 5Gemini 2.5 Pro36.2
- 6OpenAI GPT-533
- 7Gemini 3 Pro32.7
- 8Tongyi Deep Research32.1
- 9claude-sonnet-4.525.8
- 10Perplexity Sonar Deep Research24.7
| 1 | FinAgentTest | Not stated | 41.9 | Deep Research | OpenFinArena leaderboard Official board, 23 Mar 2026 |
|---|---|---|---|---|---|
| 2 | Gemini 3 Pro Deep Research | 38 | Deep Research | OpenFinArena leaderboard Official board, 30 Dec 2025 | |
| 3 | OpenAI o3-deep-research | OpenAI | 37.9 | Deep Research | OpenFinArena leaderboard Official board, 15 Oct 2025 |
| 4 | grok-4 | xAI | 37.3 | Deep Research | OpenFinArena leaderboard Official board, 15 Oct 2025 |
| 5 | Gemini 2.5 Pro | 36.2 | Deep Research | OpenFinArena leaderboard Official board, 15 Oct 2025 | |
| 6 | OpenAI GPT-5 | OpenAI | 33 | Thinking + Search | OpenFinArena leaderboard Official board, 15 Oct 2025 |
| 7 | Gemini 3 Pro | 32.7 | Thinking + Search | OpenFinArena leaderboard Official board, 30 Dec 2025 | |
| 8 | Tongyi Deep Research | Qwen | 32.1 | Deep Research | OpenFinArena leaderboard Official board, 15 Oct 2025 |
| 9 | claude-sonnet-4.5 | Anthropic | 25.8 | Thinking + Search | OpenFinArena leaderboard Official board, 15 Oct 2025 |
| 10 | Perplexity Sonar Deep Research | Perplexity | 24.7 | Deep Research | OpenFinArena leaderboard Official board, 15 Oct 2025 |