DeepSeek-V3.1 vs GPT-5 in finance

Every finance benchmark where both have a score (DeepSeek and OpenAI). Data refreshed 2 Nov 2026.

Shared benchmarks
4
DeepSeek-V3.1 only
0
GPT-5 only
3

Shared benchmarks, side by side

Each benchmark is drawn on its own scale. Bars are not comparable across rows.

  • DeepSeek-V3.1
  • GPT-5
Benchmark
DeepSeek-V3.1
GPT-5

8 benchmarks, 5 scored for both. Bold marks the better score where both have one.

AI-Tradercumulative return %, higher is better
8.4
Paper Table 1 (Reported by the authors, 1 Dec 2025)
1.6
Paper Table 1 (Reported by the authors, 1 Dec 2025)
CAIA (Crypto AI Agent Benchmark)accuracy % (majority vote, 5 runs), higher is better
49.2
Paper Tables 2-3 (Reported by the authors, 17 Jan 2026)
67.4
Paper Tables 2-3 (Reported by the authors, 17 Jan 2026)
Reasoning Models Ace the CFA Exams (CFA mock-exam evaluation)Level I accuracy %, higher is better
90.9
Paper Table 5 (Reported by the authors, 9 Dec 2025)
96.1
Paper Table 5 (Reported by the authors, 9 Dec 2025)
Reasoning Models Ace the CFA Exams (CFA mock-exam evaluation)Level III constructed-response score %, higher is better
70.8
Paper Table 5 (Reported by the authors, 9 Dec 2025)
71.8
Paper Table 5 (Reported by the authors, 9 Dec 2025)
StockBenchfinal return %, higher is better
1.1
Paper Table 2 (Reported by the authors, 2 Mar 2026)
0.300
Paper Table 2 (Reported by the authors, 2 Mar 2026)
K-FinHallubinary detection F1, higher is better
Not scored
0.823
Paper Table 3 (Reported by the authors, 28 May 2026)
LiveTradeBenchcumulative return %, higher is better
Not scored
5.3
Paper Table 2 (Reported by the authors, 5 Nov 2025)
PRBench Financeweighted rubric score (0-100), higher is better
Not scored
51.3
Scale Labs leaderboard (Official board, date not stated)

Scores from different benchmarks use different scales and are not added up. Settings can differ between sources; see each leaderboard for details.