Mistral Large (latest) vs Grok 4

Mistral Large (latest) is cheaper than Grok 4 by about 88% on a blended (75% input / 25% output) basis. Mistral Large (latest) charges $0.500000 / $1.500000 per 1M input/output tokens; Grok 4 charges $3.000000 / $15.000000.

FieldMistral Large (latest)Grok 4
Input $/Mtok$0.500000$3.000000
Output $/Mtok$1.500000$15.000000
Cache read $/Mtok$0.050000–
Cache write $/Mtok––
Input $/Mchar0.177809n/a
Output $/Mchar0.533428n/a
Context window262144256000
Tokenizermistral–

Price history

Same workload, monthly cost

Illustrative monthly cost at 10M tokens/month, split per workload preset. Not a substitute for the calculator, which uses your own volume, cache and batch assumptions.

WorkloadMistral Large (latest)Grok 4
Coding agent$5.50$36.00
RAG chat$6.50$48.00
Summarisation (batch)$6.00$42.00
Classification$5.20$32.40
Chat assistant$9.00$78.00
Content generation$13.00$126.00

Cite this page

Mistral Large (latest) vs Grok 4: $0.500000/$1.500000 vs $3.000000/$15.000000 per 1M tokens, as of 2026-10-04 (Token Barometer, https://tokenbarometer.com/compare/mistral_mistral-large-2411-vs-xai_grok-4-0709).

Mistral Large (latest) page | Grok 4 page | Open in calculator