DeepFinance benchmark: 21 real PE/IB tasks, Claude Opus 5.5 tops at 54.8%
tszzl · x · 2026-11-06
Julius Research released DeepFinance, the first ultra-long-horizon benchmark for high finance: 21 problems built from real deal data by experts from Evercore, JPM, UBS, HBS and Wharton, each averaging 27.1 hours of human work with 84 rubric criteria per task.
Leaderboard: Claude Opus 5.5 (max) leads at 54.8% ±9.1, followed by Claude Sonnet 5.5 (54.6%) and GPT-6 Astra (54.2%); GPT-6.1 Sol hits 52.9%, Gemini 4 Argon 41.3%, Kimi K3.1 Pro 34.6%, Grok 4.7 34.1%, and DeepSeek V4 Pro trails at 20.5%.
Unlike existing finance benchmarks limited to doc QA or simple model grading, DeepFinance evaluates agents end-to-end on workflows full of assumptions and conflicting evidence — and even top models only score barely above half, far from reliably diligenceing a $100M+ deal.
More from Models
- Claude Opus 5 pulls off multi-level roleplay: staying in character while knowing it's acting — repligate · 2026-11-06
- Inkling 2 Small hits 46.8% Pass@1 on APEX-Agents, up 13 points over predecessor — ziqiao_ma · 2026-11-06
- Inkling 2 Small launches with AA score of 31, weights coming soon — simonguozirui · 2026-11-06
- Inkling 2 Small matches models 8-13x pricier at just $0.34 per task — echen · 2026-11-06
- Dev complains Anthropic free credits can't be used for Code or terminal — RileyRalmuto · 2026-11-06
- Blind listening test of 62 raters ranks Gemini first, Llama last among 8 LLM composers — antiochIst · 2026-11-06