# CorpFin v2 ![Category](https://img.shields.io/badge/Category-Finance%20AI-1f6feb) ![Metric](https://img.shields.io/badge/Metric-%25%20Accuracy-2da44e) ![Saturation](https://img.shields.io/badge/Saturation-Low-d29922) ![Created](https://img.shields.io/badge/Created-2026-555) > **Quick answer:** CorpFin v2 is a second-generation corporate finance benchmark evaluating AI models on financial analysis, modeling, and task automation — the kind of work performed by corporate finance analysts. Kimi K3 scores 71.6% as of July 2026. ## At a Glance **What it tests:** A model's ability to perform accurate corporate finance analysis and modeling tasks, reflecting real financial-professional workflows. **Why it matters:** Corporate finance is a high-value, detail-sensitive domain where errors are costly. CorpFin v2 offers a domain-specific accuracy signal for finance-focused AI deployments. **Known limitations:** As a "v2" iteration, exact differences from any prior version and detailed task composition are not yet independently published outside its citation by Moonshot AI. ## What CorpFin v2 Measures CorpFin v2 evaluates a model's performance on corporate finance tasks — such as financial statement analysis, valuation modeling, and finance-domain reasoning — testing accuracy relevant to real corporate finance professional workflows. ## Benchmark Specifications | Field | Value | |---|---| | Task category | Reasoning / corporate finance | | Metric | % accuracy | | Saturation | Low | | Created by | Not yet independently documented | ## How CorpFin v2 Is Scored Models complete corporate finance analysis and modeling tasks, scored on % accuracy against expert-verified reference answers. ## State-of-the-Art Results | Rank | Model | Score | Source | Date | |---|---|---|---|---| | 1 | [Kimi K3](/models/moonshot/kimi-k3) | 71.6% | [Kimi K3 technical report](https://www.kimi.com/blog/kimi-k3) | 2026-07 | *Score sourced from Moonshot AI's Kimi K3 announcement, July 2026.* ## CorpFin v2 on Benchgen > No Benchgen results yet — [be the first to run CorpFin v2](/benchmarks/platform/corpfin-v2). ## CorpFin v2 vs Other Benchmarks | Benchmark | What it tests | Saturation | |---|---|---| | **CorpFin v2** | Corporate finance analysis & modeling | Low | | [Finance Agent v2](/benchmarks/platform/finance-agent-v2) | Financial agent task automation | Low | | [Harvey Lab-AA](/benchmarks/platform/harvey-lab-aa) | Legal knowledge-work quality | Low | | [SpreadsheetBench 2](/benchmarks/platform/spreadsheetbench-2) | Spreadsheet manipulation & formula tasks | Low | ## Run CorpFin v2 on Your Model Benchgen lets you run CorpFin v2 against your own model, tracking corporate finance analysis accuracy over time for finance-focused deployments. ## Frequently Asked Questions

What is CorpFin v2?

CorpFin v2 is a corporate finance benchmark testing AI models on financial analysis, modeling, and task automation accuracy.

What does a good score look like on CorpFin v2?

Kimi K3 reports 71.6% as of July 2026, a strong result for domain-specific corporate finance reasoning.

Who created CorpFin v2?

CorpFin v2's originating team is not yet independently documented outside of its citation in Kimi K3's July 2026 technical report.

CorpFin v2 — Results

CorpFin v2 — Phases

CorpFin v2