Finance Agent v2.1
Finance Agent v2.1, the version Vals AI now runs new models on. Run by Vals AI.
As of 24 Oct 2026, Claude Opus 5.5 leads Finance Agent v2.1 on BenchLeader with 67.8%, ahead of Gemini 4 Argon at 67.7%, across 28 model configurations with a published result.
- Published by
- Vals AI
- Category
- Agents & tools
- Index weight
- Reference only
- Models
- 28
- Data as of
- 24 Oct 2026
- Newest result
- 22 Oct 2026
Vals AI (vals.ai).
- 1Claude Opus 5.567.8%
- 2Gemini 4 Argon (high)67.7%
- 3Muse Wonder Full (max)67.6%
- 4Claude Sonnet 5.567.3%
- 5Gemini 3.8 Flash (high)65.8%
- 6Muse Wonder (max)65.5%
- 7DeepSeek V4.1 Flash (max)65.4%
- 8Claude Haiku 5.565.4%
- 9GLM 5.3 (max)65.0%
- 10MiMo-V2.6-Pro64.3%
- 11MiMo-V2.6-Flash63.3%
- 12DeepSeek V4.1 Flash 1017 (max)62.9%
- 13GPT-6.1 Sol (max)61.8%
- 14Mistral Large 4 (high)61.4%
- 15Muse Spark Open (xhigh)60.7%
28 of 28
| # | ||||
|---|---|---|---|---|
| 1 |  Claude Opus 5.5 | 67.8% | 62.0 | 2026-10-22 |
| 2 |  Gemini 4 Argonhigh | 67.7% | 70.8 | 2026-10-22 |
| 3 |  Muse Wonder Fullmax | 67.6% | 70.7 | 2026-10-22 |
| 4 | Claude Sonnet 5.5 | 67.3% | – | 2026-10-22 |
| 5 | Gemini 3.8 Flashhigh | 65.8% | 64.4 | 2026-10-22 |
| 6 | Muse Wondermax | 65.5% | 67.5 | 2026-10-22 |
| 7 |  DeepSeek V4.1 Flashmaxopen ↗ | 65.4% | 61.2 | 2026-10-22 |
| 8 | Claude Haiku 5.5 | 65.4% | – | 2026-10-22 |
| 9 |  GLM 5.3maxopen ↗ | 65.0% | 63.5 | 2026-10-22 |
| 10 |  MiMo-V2.6-Proopen ↗ | 64.3% | 64.0 | 2026-10-22 |
| 11 | MiMo-V2.6-Flashopen ↗ | 63.3% | 59.7 | 2026-10-22 |
| 12 | NewDeepSeek V4.1 Flash 1017maxopen ↗ | 62.9% | 64.0 | 2026-10-22 |
| 13 |  GPT-6.1 Solmax | 61.8% | 69.9 | 2026-10-22 |
| 14 |  Mistral Large 4highopen | 61.4% | 49.8 | 2026-10-22 |
| 15 | Muse Spark Openxhighopen ↗ | 60.7% | 62.0 | 2026-10-22 |
| 16 |  Grok 4.7xhigh | 60.6% | 61.6 | 2026-10-22 |
| 17 | GLM 5.3 Flashmaxopen ↗ | 60.6% | 54.2 | 2026-10-22 |
| 18 | GPT-6 Astramax | 60.4% | 70.4 | 2026-10-22 |
| 19 |  Kimi K3maxopen ↗ | 60.0% | 64.7 | 2026-10-22 |
| 20 |  Step 5 Previewopen | 58.9% | 60.4 | 2026-10-22 |
| 21 | New MiniMax M3.1 Flashopen ↗ | 58.8% | 62.0 | 2026-10-22 |
| 22 | New Qwen3.8 Max Internal | 58.8% | – | 2026-10-22 |
| 23 | DeepSeek V4 Promaxopen ↗ | 57.1% | 60.8 | 2026-10-22 |
| 24 | GPT-6 Lunamax | 56.8% | 60.0 | 2026-10-22 |
| 25 | New Mercury 3high | 54.7% | – | 2026-10-22 |
| 26 | MiniMax M3open ↗ | 54.3% | 56.4 | 2026-10-22 |
| 27 | New Beam 1.0open | 51.8% | 51.1 | 2026-10-22 |
| 28 | Inkling Smallopen ↗ | 44.8% | 54.8 | 2026-10-22 |
More agents & tools benchmarks · All LLM benchmarks
Cite as: BenchLeader, “Finance Agent v2.1 leaderboard”, data as of 24 Oct 2026, https://www.benchleader.com/benchmarks/vals_finance_agent_21.
Finance Agent v2.1: questions
- What does Finance Agent v2.1 measure?
- Finance Agent v2.1, the version Vals AI now runs new models on. Run by Vals AI. Scores are reported in percent of tasks solved; higher is better.
- Which AI model leads Finance Agent v2.1?
- Claude Opus 5.5 leads Finance Agent v2.1 with 67.8% as of 24 Oct 2026, ahead of Gemini 4 Argon at 67.7%.
- How many models have Finance Agent v2.1 results?
- 28 model configurations have a Finance Agent v2.1 result on BenchLeader, all taken from Vals AI.
- Who runs Finance Agent v2.1 and how often is it updated?
- Finance Agent v2.1 is published by Vals AI. BenchLeader re-reads the published results once a day and records the date each result was published.
- Does Finance Agent v2.1 count toward the BenchLeader Index?
- No. Finance Agent v2.1 is shown for reference but left out of the composite index.