DeepSeek V3.2 — pricing, providers, and benchmarks

Parameters
671B
Context window
131k tokens
License
deepseek
Released
2025-05-07

DeepSeek V3.2 is the latest in DeepSeek's V3 mixture-of-experts series, with 671B total parameters but only 37B active per forward pass — meaning hosting providers can serve it at prices closer to a 70B dense model while delivering 671B-class quality on many tasks. Released in late 2025, V3.2 added improved multilingual support and a 30% reduction in inference cost over V3 according to DeepSeek's release notes. It punches well above its weight for general-purpose chat and coding, and is the cheapest frontier-class open-weights model currently hosted. Input pricing across major providers sits in the $0.20–$0.27/1M range; output is more expensive ($0.85–$1.10/1M) due to the MoE routing overhead during generation.

Provider pricing

Sorted by total monthly cost for 100M input + 10M output tokens.

ProviderInput / 1MOutput / 1MMonthly costContext
Together AI$0.2700$1.1000$38.00131k

Frequently asked questions

How much does it cost to run DeepSeek V3.2 for 100M tokens?

Running DeepSeek V3.2 with 100M input and 10M output tokens per month costs approximately $38.00 on Together AI, the cheapest available provider as of the latest pricing data. Costs vary significantly depending on your input/output ratio and whether you use prompt caching.

What is the cheapest provider for DeepSeek V3.2?

Together AI currently offers DeepSeek V3.2 at the lowest total cost for a standard workload. Prices change frequently — check the table above for the latest data.

What context window does DeepSeek V3.2 support?

DeepSeek V3.2 supports a context window of 131,072 tokens. Individual providers may cap this lower — see the pricing table for per-provider context limits.