Codestral 2 vs DeepSeek V4

DeepSeek V4 is cheaper on input ($0.40 vs $0.50 per 1M tokens).

SpecCodestral 2DeepSeek V4
ProviderMistralDeepSeek
ModalityTextText
Context window256K256K
Input / 1M$0.50$0.40best
Output / 1M$1.50$1.20best
ReleasedNov 2025Jan 2026
Best forCode generation, High-volume / cheapCode generation, Chat / Assistant, High-volume / cheap

FAQ

Is Codestral 2 cheaper than DeepSeek V4?
DeepSeek V4 is cheaper, at $0.40 per 1M input tokens versus $0.50.
Which has the larger context window, Codestral 2 or DeepSeek V4?
Both support 256K tokens of context.

More comparisons