AI Hub
← All models

Codestral vs DeepSeek V3.2 Speciale

Mistral AI vs DeepSeek — benchmarks, pricing, and capabilities side by side.

  • DeepSeek V3.2 Speciale is cheaper ($0.29 vs $0.30 per 1M input)
  • Codestral has a larger context window (256K)
CodestralDeepSeek V3.2 Speciale
Intelligence index89.9
DeveloperMistral AIDeepSeek
TypeLLMLLM
AccessAPI onlyOpen weights
Context window256,000 tokens163,840 tokens
Input price$0.30 / 1M$0.29 / 1M
Output price$0.90 / 1M$0.43 / 1M
Speed
ReleasedAugust 1, 2025December 1, 2025
Parameters
Input modalitiesTextText
Output modalitiesTextText