deepseek/deepseek-v3.2DeepSeek-V3.2 is DeepSeek’s open-weights AI model, released Dec 1, 2025. It costs $0.25/M input and $0.42/M output tokens, and handles a 164K-token context window.
DeepSeek-V3.2 is a large language model designed to harmonize high computational efficiency with strong reasoning and agentic tool-use performance.
Scores on standardized evaluations. Higher is better — and rank shows where DeepSeek-V3.2 lands among all models tracked on Model Beat.
Abstract visual reasoning
Harder abstract reasoning
Graduate-level scientific reasoning
Frontier of human expert knowledge
Broad expert knowledge
Common-sense trick questions
Factual accuracy & hallucination
Novel ML problem-solving
Contamination-free coding
Scientific research coding
Human-rated web development
Olympiad-qualifier math
Research-level math problems
Hardest research math
Multi-step agentic tasks
Command-line agentic tasks
Tool-agent-user reliability
DeepSeek-V3.2 is a large language model designed to harmonize high computational efficiency with strong reasoning and agentic tool-use performance.
DeepSeek-V3.2 is an AI model developed by DeepSeek, part of the DeepSeek family, released Dec 1, 2025.
DeepSeek-V3.2 is served from $0.25 per million input tokens and $0.42 per million output tokens (cheapest credible provider on OpenRouter: SiliconFlow).
DeepSeek-V3.2 supports a context window of up to 164K tokens.
On standardized evaluations tracked by Epoch AI, DeepSeek-V3.2 scores 86.2% on LiveCodeBench, 86.2% on MMLU-Pro, 90.6% on τ²-bench.
DeepSeek-V3.2 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.