ModelBeat
Models/DeepSeek-V3.2-Exp
All models
D

DeepSeek: DeepSeek-V3.2-Exp

deepseek/deepseek-v3.2-exp

DeepSeek-V3.2-Exp is DeepSeek’s open-weights AI model, released Sep 29, 2025. It costs $0.27/M input and $0.41/M output tokens, and handles a 164K-token context window.

DeepSeek-V3.2-Exp is an experimental large language model released by DeepSeek as an intermediate step between V3.1 and future architectures.

Context
164K
Input
$0.27 / 1M
Output
$0.41 / 1M
Released
Sep 29, 2025
Modalities
texttext

Benchmarks

Scores on standardized evaluations. Higher is better — and rank shows where DeepSeek-V3.2-Exp lands among all models tracked on Model Beat.

31.0
Intelligence Index
Epoch AI
31st percentile of tracked models
35.0
Coding Index
Epoch AI
35th percentile of tracked models
9.0
Agentic Index
Epoch AI
9th percentile of tracked models

Reasoning

4 evals
GPQA Diamond79.7%

Graduate-level scientific reasoning

Humanity's Last Exam13.8%

Frontier of human expert knowledge

MMLU-Pro85.0%

Broad expert knowledge

WeirdML39.5%

Novel ML problem-solving

Coding

3 evals
LiveCodeBench78.9%

Contamination-free coding

SciCode37.7%

Scientific research coding

WebDev Arena1367

Human-rated web development

Math

1 evals
AIME 2024/202587.7%

Olympiad-qualifier math

Agentic & Tools

1 evals
τ²-bench33.9%

Tool-agent-user reliability

Compare DeepSeek-V3.2-Exp with

Frequently asked questions

What is DeepSeek-V3.2-Exp?

DeepSeek-V3.2-Exp is an experimental large language model released by DeepSeek as an intermediate step between V3.1 and future architectures.

Who created DeepSeek-V3.2-Exp?

DeepSeek-V3.2-Exp is an AI model developed by DeepSeek, part of the DeepSeek family, released Sep 29, 2025.

How much does DeepSeek-V3.2-Exp cost?

DeepSeek-V3.2-Exp is served from $0.27 per million input tokens and $0.41 per million output tokens (cheapest credible provider on OpenRouter: Novita).

What is DeepSeek-V3.2-Exp's context window?

DeepSeek-V3.2-Exp supports a context window of up to 164K tokens.

How does DeepSeek-V3.2-Exp perform on benchmarks?

On standardized evaluations tracked by Epoch AI, DeepSeek-V3.2-Exp scores 85.0% on MMLU-Pro, 78.9% on LiveCodeBench, 87.7% on AIME 2024/2025.

Is DeepSeek-V3.2-Exp open source?

DeepSeek-V3.2-Exp is released with open weights (Open weights (unrestricted)), so the model can be downloaded and self-hosted.

Benchmark data from Epoch AI (CC BY) and Artificial Analysis; pricing, specs & descriptions from OpenRouter. Data updated Jul 30, 2026.