AI Benchmark Model

DeepSeek V3.1 (Reasoning)

Benchmark scores, pricing, speed, and model comparisons for DeepSeek V3.1 (Reasoning).

Key scores

Intelligence13.5Index score
Math89.7Index score
Blended Cost$0.86Per 1M tokens

Review benchmark scores, pricing, performance data, and generated comparisons for this AI model.

Benchmark results

9 tracked benchmark rows for DeepSeek V3.1 (Reasoning).

AIME 2025 Math
89.7%
MMLU Pro Knowledge
85.1%
LiveCodeBench Code
78.4%
GPQA Reasoning
77.9%
LCR Other
56.7%
IFBench Other
41.5%
TAU2 Other
37.4%
TerminalBench Hard Other
25.0%
HLE Other
14.3%