# Finance Agent v2

> Vals AI benchmark for realistic financial analyst agent tasks across qualitative analysis, quantitative analysis, market work, comparables, precedents, earnings, disclosure, and modeling.

Canonical page: https://benchlm.ai/benchmarks/financeagentv2

- Category: [Agentic](/agentic)
- Last updated: September 21, 2026

## About Finance Agent v2

- Year: 2026
- Tasks: Financial analyst task categories
- Format: Mean score across repeated runs
- Difficulty: Professional expert-task agent workflow
- Paper: [Finance Agent v2](https://www.vals.ai/benchmarks/fabv2)

Vals reports Finance Agent v2 as a multi-category benchmark with severity-weighted partial credit and repeated runs per model. BenchLM mirrors the public Vals leaderboard as a display-only expert-task benchmark.

Finance Agent v2 is currently displayed on BenchLM for reference, but it is excluded from the weighted scoring formula.

## Leaderboard (63 models)

| Rank | Model | Configuration | Creator | Score |
|------|-------|---------------|---------|-------|
| 1 | [Gemini 3.8 Flash](/models/gemini-3-8-flash) | high reasoning | Google | 61.4% |
| 2 | [Muse Spark 1.2](/models/muse-spark-1-2) | xhigh reasoning | Meta | 60.6% |
| 3 | [Muse Spark 1.3 Max](https://www.vals.ai/models/meta_muse_spark_1_3_max) | max reasoning | Meta | 60.0% |
| 4 | [Gemini 3.7 Flash](/models/gemini-3-7-flash) | high reasoning | Google | 59.0% |
| 5 | [Muse Spark 1.3](/models/muse-spark-1-3) | xhigh reasoning | Meta | 58.9% |
| 6 | [Claude Fable 5.1](/models/claude-fable-5-1) | — | Anthropic | 58.9% |
| 7 | [Claude Opus 5](/models/claude-opus-5) | — | Anthropic | 58.6% |
| 8 | [Claude Opus 5.5](/models/claude-opus-5-5) | — | Anthropic | 58.6% |
| 9 | [Gemini 3.5 Flash](/models/gemini-3-5-flash) | high reasoning | Google | 57.9% |
| 10 | [GLM-5.3-Flash](/models/glm-5-3-flash) | max reasoning | Z.AI | 57.9% |
| 11 | [Muse Spark 1.1](/models/muse-spark-1-1) | xhigh reasoning | Meta | 57.2% |
| 12 | [Claude Fable 5](/models/claude-fable) | — | Anthropic | 56.3% |
| 13 | [Gemini 3.6 Flash](/models/gemini-3-6-flash) | high reasoning | Google | 56.3% |
| 14 | [GLM-5.3](/models/glm-5-3) | max reasoning | Z.AI | 55.8% |
| 15 | [Hy4 preview](/models/hy4-preview) | — | Tencent | 55.1% |
| 16 | [GPT-5.6 Luna](/models/gpt-5-6-luna) | max reasoning | OpenAI | 55.0% |
| 17 | [GPT-5.6 Terra](/models/gpt-5-6-terra) | max reasoning | OpenAI | 54.4% |
| 18 | [Kimi K3](/models/kimi-k3) | — | Moonshot AI | 54.4% |
| 19 | [Claude Opus 4.8](/models/claude-opus-4-8) | — | Anthropic | 53.9% |
| 20 | [Claude Sonnet 5](/models/claude-sonnet-5) | — | Anthropic | 53.9% |
| 21 | [GPT-5.6 Sol](/models/gpt-5-6-sol) | max reasoning | OpenAI | 53.8% |
| 22 | [Grok 4.6](/models/grok-4-6) | high reasoning | xAI | 53.7% |
| 23 | [GPT-6 Astra](/models/gpt-6-astra) | max reasoning | OpenAI | 53.5% |
| 24 | [DeepSeek V4.1 Flash](/models/deepseek-v4-1-flash) | high reasoning | DeepSeek | 53.5% |
| 25 | [Grok 4.7](/models/grok-4-7) | xhigh reasoning | xAI | 52.3% |
| 26 | [GPT-5.5](/models/gpt-5-5) | xhigh reasoning | OpenAI | 51.8% |
| 27 | [Claude Opus 4.7](/models/claude-opus-4-7) | — | Anthropic | 51.5% |
| 28 | [Claude Sonnet 4.6](/models/claude-sonnet-4-6) | — | Anthropic | 51.0% |
| 29 | [Qwen3.8 Max](/models/qwen3-8-max) | — | Alibaba | 50.6% |
| 30 | [DeepSeek V4 Pro 0813](/models/deepseek-v4-pro-0813) | max reasoning | DeepSeek | 50.4% |
| 31 | [GLM-5.2](/models/glm-5-2) | — | Z.AI | 49.7% |
| 32 | [DeepSeek V4 Flash 0731](/models/deepseek-v4-flash-0731) | high reasoning | DeepSeek | 49.5% |
| 33 | [Qwen3.8-27B](/models/qwen3-8-27b) | xhigh reasoning | Alibaba | 48.6% |
| 34 | [Grok 4.5](/models/grok-4-5) | high reasoning | xAI | 48.3% |
| 35 | [MiniMax M3](/models/minimax-m3) | — | MiniMax | 48.3% |
| 36 | [Qwen3.7 Max](/models/qwen3-7-max) | — | Alibaba | 47.8% |
| 37 | [Gemini 3.5 Flash-Lite](/models/gemini-3-5-flash-lite) | high reasoning | Google | 47.4% |
| 38 | [Inkling](/models/inkling) | 0.99 reasoning | Thinking Machines Lab | 46.6% |
| 39 | [GPT-5.4 mini](/models/gpt-5-4-mini) | xhigh reasoning | OpenAI | 45.4% |
| 40 | [Kimi K2.6](/models/kimi-2-6) | — | Moonshot AI | 44.9% |
| 41 | [GLM-5.1](/models/glm-5-1) | — | Z.AI | 44.8% |
| 42 | [DeepSeek V4 Pro 0813](/models/deepseek-v4-pro-0813) | max reasoning | DeepSeek | 44.1% |
| 43 | [Gemini 3.1 Pro Preview](https://www.vals.ai/models/google_gemini-3.1-pro-preview) | high reasoning | Google | 43.0% |
| 44 | [Gemini 3 Flash Preview](https://www.vals.ai/models/google_gemini-3-flash-preview) | high reasoning | Google | 42.6% |
| 45 | [MiMo-V2.5-Pro](/models/mimo-v2-5-pro) | — | Xiaomi | 41.5% |
| 46 | [Inkling-Small](/models/inkling-small) | 0.99 reasoning | Thinking Machines Lab | 41.3% |
| 47 | [Qwen3.6 Plus](/models/qwen3-6-plus) | — | Alibaba | 40.8% |
| 48 | [Qwen3.7 Plus](/models/qwen3-7-plus) | — | Alibaba | 38.2% |
| 49 | [GPT-5.4 nano](/models/gpt-5-4-nano) | high reasoning | OpenAI | 38.2% |
| 50 | [Grok 4.3](/models/grok-4-3) | high reasoning | xAI | 37.7% |
| 51 | [Nemotron 3 Ultra 550b A55b](https://www.vals.ai/models/nvidia_nemotron-3-ultra-550b-a55b) | — | Nvidia | 37.7% |
| 52 | [MiMo-V2.5](/models/mimo-v2-5) | — | Xiaomi | 36.7% |
| 53 | [Kimi K2.5 Thinking](https://www.vals.ai/models/kimi_kimi-k2.5-thinking) | — | Moonshot AI | 35.8% |
| 54 | [Mistral Medium 3.5](https://www.vals.ai/models/mistralai_mistral-medium-3.5) | high reasoning | Mistral AI | 32.1% |
| 55 | [Claude Haiku 4.5 Thinking](/models/claude-haiku-4-5-thinking) | — | Anthropic | 31.0% |
| 56 | [Ling 3.0 Flash 2607](https://www.vals.ai/models/ant_ling-3.0-flash-2607) | — | Ant | 30.3% |
| 57 | [Gemini 3.1 Flash Lite Preview](https://www.vals.ai/models/google_gemini-3.1-flash-lite-preview) | high reasoning | Google | 30.0% |
| 58 | [Grok 4.20 0309 Reasoning](https://www.vals.ai/models/grok_grok-4.20-0309-reasoning) | — | xAI | 28.5% |
| 59 | [MiniMax M2.7](/models/minimax-m2-7) | — | MiniMax | 27.9% |
| 60 | [Laguna M.1](/models/laguna-m-1) | — | Poolside | 25.0% |
| 61 | [Mercury 2.5](/models/mercury-2-5) | high reasoning | Inception | 18.6% |
| 62 | [Nemotron Lightning 3p5 30b A3b](https://www.vals.ai/models/fireworks_nemotron-lightning-3p5-30b-a3b) | — | Fireworks AI | 18.5% |
| 63 | [Laguna XS.2](/models/laguna-xs-2) | — | Poolside | 15.6% |

## FAQ

### What does Finance Agent v2 measure?

Vals AI benchmark for realistic financial analyst agent tasks across qualitative analysis, quantitative analysis, market work, comparables, precedents, earnings, disclosure, and modeling.

### Which model leads the published Finance Agent v2 snapshot?

Gemini 3.8 Flash currently leads the published Finance Agent v2 snapshot with a score of 61.4%.

### How many models are evaluated on Finance Agent v2?

The September 21, 2026 contains 63 AI models.

### Does Finance Agent v2 affect BenchLM's overall score?

Not directly. Finance Agent v2 is still displayed on BenchLM for reference, but it is excluded from the weighted scoring formula.