Learning from real decisionsDissei / Data

Teaching models
financial judgment.

Carefully built training data, evaluations and environments, grounded in real financial work, for AI labs and enterprise teams.

For labs For enterprise
Financial AI reasoning leaderboard Training data Evaluations RL environments How we build and anonymize data

Looking for financing or a lending partnership? Visit Dissei Lending.

Learning, under examination.

Studies in what models learn from finance, and where their judgment breaks.

Practitioner guides
Latest researchAll research
012026-09-25MethodsJev’s probabilities, put to the test022026-09-17Research noteOnly if you ask: planting wrong numbers in a 20-step credit worksheet with three models032026-09-10EvalsSaying 70, meaning 40: guessing probabilities with five mid-tier models
Financial AI Leaderboard

Financial reasoning,
measured.

How AI models and LLMs reason from financial evidence, compared across seven reasoning categories and scored on a 0–100 continuous reward scale, not accuracy.

View the financial reasoning leaderboard Inspect the public dataset

Pooled evaluation scores by model

Historical snapshot · 2026-08-31

Observed score / 100

Source: Dissei Leaderboard

Models were evaluated on differing cases, not matched head-to-head. Compare figures with care.

View data table
Pooled evaluation scores by model
ModelScore / 100
Claude Opus 4.844.26
DeepSeek v4-flash25.49
GPT-5.6 Sol45.53
Kimi K339.76
Muse Spark 1.242.41