D
DeepSeek V3.2OPEN
DeepSeek-V3.2 is a large language model designed to harmonize high computational efficiency with strong reasoning and agentic tool-use performance.
Prices update automatically — checked daily against provider list prices.
- $0.29 per 1M tokens, assuming 3 input for 1 output the convention used by Artificial Analysis
- cheaper than 55% of the 191 open-weight models tracked here
- 3 of 14 DeepSeek models tracked here, from cheapest
Benchmarks & performance
Independent benchmark scores for DeepSeek V3.2, measured by Artificial Analysis. Higher is better. Measurement mode: Non-reasoning.
For the same budget, DeepSeek V3.2 delivers 13 times more capability. Open weights also mean you can self-host it and pay nothing per token.
Models at this level cost $0.96 per 1M tokens (median of 21) — this one costs $0.29.
Cheaper and better on this index: GLM 5.3 Flash · GPT-5.6 Luna (batch) · DeepSeek V4 Flash 0731 and 12 more
About this model
DeepSeek V3.2 is an open-weight AI model by DeepSeek. You can download and self-host it for free; the prices below are hosted-API list prices, tracked daily, for when you prefer convenience over self-hosting.
Frequently asked questions
What is DeepSeek V3.2?
DeepSeek V3.2 is an AI language model from DeepSeek. It is open-weight: you can download it and run it on your own hardware, for free. It scores 25.1 on the Artificial Analysis intelligence index.
Is DeepSeek V3.2 free?
The weights are free and open — you can self-host DeepSeek V3.2 and pay nothing per token. If you prefer a hosted API, list prices are $0.26 per million input tokens and $0.38 per million output tokens.
What is DeepSeek V3.2 good at?
Independent benchmarks from Artificial Analysis give it GPQA 75.1%, MMLU-Pro 83.7%, Humanity's Last Exam 11.2%, Long Context Reasoning 42.7%, LiveCodeBench 59.3%, SciCode 38.7%, AIME 2025 59%, IFBench 49%, τ²-Bench 78.9%, Terminal-Bench Hard 32.6%. It is particularly used for mathematical reasoning.
Can I self-host DeepSeek V3.2?
Yes. DeepSeek V3.2 has open weights, so you can download it and run it on your own GPU or server with tools like Ollama, vLLM or llama.cpp — with no per-token cost.