MacroLens - Private Company Valuation: leaderboard
Zero-shot valuation without market prices (MacroLens, Trirat et al., 2026): the same 1,324-ticker holdout as the public valuation task, with every price-derived input and the derived valuation-ratio block removed, approximating an unlisted firm. Median absolute percentage error of the market-capitalization estimate in percent, per-instance errors capped at 1,000%, for LLMs run with reasoning disabled; lower is better.
Metric: Median absolute percentage error (%; market-capitalization estimate for the 1,324-ticker holdout with every price-derived input and the derived valuation-ratio block removed; per-instance APE capped at 1,000%; daily granularity; zero-shot LLM with reasoning disabled). Source: arxiv.org. Saturation forecast: Estimated already saturated. 4 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.1 (Non-reasoning) | 66.23 | #225 (GPT-5.1) |
| 2 | Llama 4 Scout | 68.92 | #1787 |
| 3 | Qwen 3.5 27B (Non-reasoning) | 73.42 | #358 (Qwen 3.5 27B) |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=macrolens-private-company-valuation · How It Works · Data refreshed daily, snapshot 2026-11-12.