# Vals EMB (EMB)

> Evaluating agents on Excel-based financial modeling tasks

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

- Category: [Agentic](/agentic)
- Last updated: October 12, 2026

Evaluation results published by [Vals AI](https://www.vals.ai/benchmarks/emb). Published leaderboard aggregates and evaluation settings only. Dataset contents are not included.

## About EMB

- Year: 2026
- Tasks: Excel-based financial modeling tasks
- Format: Accuracy score
- Difficulty: Professional finance modeling
- Paper: [EMB](https://www.vals.ai/benchmarks/emb)

BenchLM mirrors the public Vals AI EMB leaderboard as display-only external evidence. The captured snapshot preserves overall scores, task-level scores where Vals publishes them, uncertainty, latency, and cost-per-test metadata. It is excluded from BenchLM weighted rankings.

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

## Leaderboard (74 models)

| Rank | Model | Configuration | Creator | Score |
|------|-------|---------------|---------|-------|
| 1 | [Claude Fable 5.1](/models/claude-fable-5-1) | — | Anthropic | 76.67% |
| 2 | [Claude Opus 5.5](/models/claude-opus-5-5) | — | Anthropic | 75.94% |
| 3 | [Claude Sonnet 5.5](/models/claude-sonnet-5-5) | — | Anthropic | 75.71% |
| 4 | [Gemini 4 Argon](/models/gemini-4-argon) | high reasoning | Google | 75.24% |
| 5 | [Claude Fable 5](/models/claude-fable) | — | Anthropic | 73.67% |
| 6 | [Claude Opus 5](/models/claude-opus-5) | — | Anthropic | 73.56% |
| 7 | [GPT-5.6 Sol](/models/gpt-5-6-sol) | max reasoning | OpenAI | 72.34% |
| 8 | [Gemini 3.8 Flash](/models/gemini-3-8-flash) | high reasoning | Google | 72.20% |
| 9 | [Muse Wonder Full](/models/muse-wonder-full) | max reasoning | Meta | 71.92% |
| 10 | [GPT-6 Astra](/models/gpt-6-astra) | max reasoning | OpenAI | 71.70% |
| 11 | [GPT-6 Sol](/models/gpt-6-sol) | max reasoning | OpenAI | 71.53% |
| 12 | [Gemini 3.7 Flash](/models/gemini-3-7-flash) | high reasoning | Google | 71.33% |
| 13 | [GPT-6.1 Sol](/models/gpt-6-1-sol) | max reasoning | OpenAI | 70.77% |
| 14 | [Claude Opus 4.8](/models/claude-opus-4-8) | — | Anthropic | 69.37% |
| 15 | [GPT-6 Luna](/models/gpt-6-luna) | max reasoning | OpenAI | 68.52% |
| 16 | [Muse Spark 1.3 Max](https://www.vals.ai/models/meta_muse_spark_1_3_max) | max reasoning | Meta | 67.43% |
| 17 | [GPT-5.6 Luna](/models/gpt-5-6-luna) | max reasoning | OpenAI | 67.12% |
| 18 | [Muse Wonder](/models/muse-wonder) | max reasoning | Meta | 67.08% |
| 19 | [Grok 4.7](/models/grok-4-7) | xhigh reasoning | xAI | 66.99% |
| 20 | [Kimi K3](/models/kimi-k3) | max reasoning | Moonshot AI | 66.68% |
| 21 | [Claude Sonnet 5](/models/claude-sonnet-5) | — | Anthropic | 66.32% |
| 22 | [GPT-5.6 Terra](/models/gpt-5-6-terra) | max reasoning | OpenAI | 66.20% |
| 23 | [MiMo-V2.6-Flash](/models/mimo-v2-6-flash) | — | Xiaomi | 65.46% |
| 24 | [Gemini 3.6 Flash](/models/gemini-3-6-flash) | high reasoning | Google | 65.41% |
| 25 | [GPT-5.5](/models/gpt-5-5) | xhigh reasoning | OpenAI | 64.54% |
| 26 | [Claude Opus 4.7](/models/claude-opus-4-7) | — | Anthropic | 63.73% |
| 27 | [Gemini 3.5 Flash](/models/gemini-3-5-flash) | high reasoning | Google | 63.55% |
| 28 | [Ember-1](/models/ember-1) | — | Fireworks | 63.37% |
| 29 | [MiMo-V2.6-Pro](/models/mimo-v2-6-pro) | — | Xiaomi | 62.86% |
| 30 | [Grok 4.6](/models/grok-4-6) | high reasoning | xAI | 62.73% |
| 31 | [Muse Spark 1.3](/models/muse-spark-1-3) | xhigh reasoning | Meta | 62.71% |
| 32 | [GLM-5.2](/models/glm-5-2) | — | Z.AI | 61.53% |
| 33 | [Claude Sonnet 4.6](/models/claude-sonnet-4-6) | — | Anthropic | 60.15% |
| 34 | [Qwen3.8 Max](/models/qwen3-8-max) | — | Alibaba | 60.07% |
| 35 | [Qwen3.8-27B](/models/qwen3-8-27b) | xhigh reasoning | Alibaba | 59.66% |
| 36 | [Claude Haiku 5.5](/models/claude-haiku-5-5) | — | Anthropic | 59.02% |
| 37 | [Hy4 preview](/models/hy4-preview) | — | Tencent | 58.08% |
| 38 | [Kimi K2.6](/models/kimi-2-6) | — | Moonshot AI | 57.85% |
| 39 | [DeepSeek V4.1 Flash 1017](/models/deepseek-v4-1-flash) | high reasoning | DeepSeek | 57.21% |
| 40 | [Step 5 Preview](/models/step-5-preview) | — | StepFun | 57.12% |
| 41 | [Qwen3.7 Max](/models/qwen3-7-max) | — | Alibaba | 57.04% |
| 42 | [Muse Spark 1.2](/models/muse-spark-1-2) | xhigh reasoning | Meta | 56.98% |
| 43 | [DeepSeek V4 Flash 0731](/models/deepseek-v4-flash-0731) | high reasoning | DeepSeek | 56.98% |
| 44 | [Muse Spark 1.1](/models/muse-spark-1-1) | xhigh reasoning | Meta | 56.36% |
| 45 | [GLM-5.3](/models/glm-5-3) | max reasoning | Z.AI | 56.34% |
| 46 | [Mistral Large 4](/models/mistral-large-4) | high reasoning | Mistral | 56.21% |
| 47 | [GLM-5.3-Flash](/models/glm-5-3-flash) | max reasoning | Z.AI | 55.93% |
| 48 | [MiMo-V2.5-Pro](/models/mimo-v2-5-pro) | — | Xiaomi | 55.23% |
| 49 | [MiMo-V2.5](/models/mimo-v2-5) | — | Xiaomi | 55.09% |
| 50 | [Grok 4.5](/models/grok-4-5) | high reasoning | xAI | 52.91% |
| 51 | [DeepSeek V4 Pro 0813](/models/deepseek-v4-pro-0813) | max reasoning | DeepSeek | 52.80% |
| 52 | [Gemini 3.1 Pro Preview](https://www.vals.ai/models/google_gemini-3.1-pro-preview) | high reasoning | Google | 52.62% |
| 53 | [DeepSeek V4 Pro 0813](/models/deepseek-v4-pro-0813) | max reasoning | DeepSeek | 51.62% |
| 54 | [Qwen3.7 Plus](/models/qwen3-7-plus) | — | Alibaba | 49.34% |
| 55 | [MiniMax M3](/models/minimax-m3) | — | MiniMax | 47.76% |
| 56 | [GPT-5.4 mini](/models/gpt-5-4-mini) | xhigh reasoning | OpenAI | 45.43% |
| 57 | [GPT-5.4 nano](/models/gpt-5-4-nano) | high reasoning | OpenAI | 44.75% |
| 58 | [Gemini 3.5 Flash-Lite](/models/gemini-3-5-flash-lite) | high reasoning | Google | 43.23% |
| 59 | [Inkling](/models/inkling) | 0.99 reasoning | Thinking Machines Lab | 38.36% |
| 60 | [Ling 3.0 Flash Af Rc3](https://www.vals.ai/models/ant_ling-3.0-flash-af-rc3) | — | Ant | 36.83% |
| 61 | [Qwen3.6 Plus](/models/qwen3-6-plus) | — | Alibaba | 32.87% |
| 62 | [Inkling-Small](/models/inkling-small) | 0.99 reasoning | Thinking Machines Lab | 31.76% |
| 63 | [Nemotron 3 Ultra 550b A55b](https://www.vals.ai/models/nvidia_nemotron-3-ultra-550b-a55b) | — | Nvidia | 31.73% |
| 64 | [Kimi K2.5 Thinking](https://www.vals.ai/models/kimi_kimi-k2.5-thinking) | — | Moonshot AI | 28.47% |
| 65 | [MiniMax M2.7](/models/minimax-m2-7) | — | MiniMax | 28.45% |
| 66 | [Gemini 3 Flash Preview](https://www.vals.ai/models/google_gemini-3-flash-preview) | high reasoning | Google | 26.02% |
| 67 | [Ling 3.0 Flash 2607](https://www.vals.ai/models/ant_ling-3.0-flash-2607) | — | Ant | 25.01% |
| 68 | [Claude Haiku 4.5 Thinking](/models/claude-haiku-4-5-thinking) | — | Anthropic | 22.66% |
| 69 | [Grok 4.3](/models/grok-4-3) | high reasoning | xAI | 18.39% |
| 70 | [Nemotron Lightning 3p5 30b A3b](https://www.vals.ai/models/fireworks_nemotron-lightning-3p5-30b-a3b) | — | Fireworks AI | 16.99% |
| 71 | [Grok 4.20 0309 Reasoning](https://www.vals.ai/models/grok_grok-4.20-0309-reasoning) | — | xAI | 11.74% |
| 72 | [Mistral Medium 3.5](https://www.vals.ai/models/mistralai_mistral-medium-3.5) | high reasoning | Mistral AI | 11.19% |
| 73 | [Mercury 2.5](/models/mercury-2-5) | high reasoning | Inception | 8.79% |
| 74 | [Gemini 3.1 Flash Lite Preview](https://www.vals.ai/models/google_gemini-3.1-flash-lite-preview) | high reasoning | Google | 8.63% |

## FAQ

### What does EMB measure?

Evaluating agents on Excel-based financial modeling tasks

### Which model leads the published EMB snapshot?

Claude Fable 5.1 currently leads the published EMB snapshot with a score of 76.67%.

### How many models are evaluated on EMB?

The October 12, 2026 contains 74 AI models.

### Does EMB affect BenchLM's overall score?

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