FINESSE-Bench - Trading and Technical Analysis: leaderboard

FINESSE-Bench Trading and Technical Analysis covers the applied group of FINESSE-Bench (2026): derivatives trading, trading-context technical analysis and CFTe-like technical-analysis questions, 1,738 in all. Answers are marked correct by a GPT-5.2 judge; the score is accuracy in percent weighted by dataset size, higher is better.

Metric: Accuracy (%) over the trading and technical-analysis group of FINESSE-Bench: Trading_derivatives (544), Trading_TA (413) and CFTe-like Level 1 (781), the per-dataset accuracies weighted by question count; a GPT-5.2 judge marks each answer correct or incorrect against the reference; zero-shot, one fixed prompt per task type, temperature 0 where possible, reasoning configurations with medium effort where the model offers them; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 31 models tracked. Updated 7 Oct 2026.

Top models

#ModelSettingScoreOverall rank
1Kimi K2.584.98#266
2Qwen 3.5 Plus (2026-02-15)84.75#194
3GPT-5.2Medium84.24#185
4Qwen 3.5 397B A17B83.31#264
5Qwen 3.5 122B A10B82.79#346
6Claude Sonnet 4.6Medium82.45#173
7GLM-4.782.05#380
8GLM-581.3#269
9Qwen 3.5 27B80.38#375
10Qwen 3 235B A22B 2507 (Thinking)80.27#466
11GPT-5.4Medium79.69#135
12Qwen 3.5 Flash (02-23)78.94#557
13Qwen 3.5 35B A3B78.83#492
14Claude 3.7 SonnetThinking78.54#470
15DeepSeek R1 052878.48#428

No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=finesse-bench-trading-and-technical-analysis · How It Works · Data refreshed daily, snapshot 2026-11-21.