Mistral AI

Mistral API Pricing 2026

Complete cost breakdown for Mistral Large, Small, Nemo, and Codestral — with comparisons to Claude, GPT-4o, and Gemini.

Mistral Model Pricing Table

All prices are per million tokens (MTok). Mistral bills input and output tokens separately.

ModelInput (per MTok)Output (per MTok)ContextBest For
Mistral Large$2.00$6.00128KComplex reasoning, multilingual
Mistral Small$0.20$0.6032KSimple tasks, classification
Mistral Nemo$0.15$0.15128KBudget-tier inference, RAG
Codestral$0.20$0.6032KCode generation, autocomplete
Mistral Embed$0.108KEmbeddings, semantic search

Mistral vs Claude vs GPT-4o — Head-to-Head

Comparing flagship models and their budget tiers across all major providers.

ModelInput (MTok)Output (MTok)Cache DiscountContext
Mistral Nemo$0.15$0.15None128K
GPT-4o-mini$0.15$0.6050%128K
Claude Haiku 4.5$0.25$1.2590%200K
Mistral Small$0.20$0.60None32K
Mistral Large$2.00$6.00None128K
GPT-4o$2.50$10.0050%128K
Claude Sonnet 4.6$3.00$15.0090%200K
Gemini 1.5 Pro$3.50$10.50None1M
Claude Opus 4.7$15.00$75.0090%200K
The caching caveat: Mistral's headline prices look great, but it offers no prompt caching. Claude Sonnet's 90% cache-read discount ($0.30/MTok) means a system prompt queried 10,000 times costs $3 cached vs $20,000 uncached at Mistral Large rates. For high-volume apps with repeated context, Claude's effective cost often beats Mistral. Calculate your specific workload before committing.

When to Choose Mistral

European Compliance

GDPR

Mistral is a French company with EU data residency options. Ideal for EU-regulated workloads where data locality matters.

Multilingual Apps

10+ langs

Mistral Large has particularly strong French, German, Spanish, Italian performance — better than GPT-4o-mini for EU language tasks.

Budget Inference

$0.15/MTok

Mistral Nemo at $0.15/MTok (both input and output) is one of the cheapest capable models available for non-critical tasks.

Code Generation

Codestral

At $0.20/MTok, Codestral is 12.5× cheaper than Claude Sonnet for autocomplete and boilerplate-heavy coding pipelines.

Monthly Cost Example: 100M Tokens/Month

ModelInput (50M tokens)Output (50M tokens)Monthly Total
Mistral Nemo$7.50$7.50$15
Mistral Small$10$30$40
Claude Haiku (cached 90%)$1.25 cached$62.50~$64
Mistral Large$100$300$400
GPT-4o$125$500$625
Claude Sonnet (cached 90%)$15 cached$750~$765

Mistral vs Claude: Which Should You Choose?

FactorMistral LargeClaude Sonnet 4.6
Input price$2.00/MTok (cheaper)$3.00/MTok
Output price$6.00/MTok (cheaper)$15.00/MTok
Prompt cachingNot available90% discount
Context window128K200K
Reasoning qualityGoodExcellent
EU data residencyAvailableUS-based
Tool use / function callingGoodBest-in-class
Open weights optionYes (Mistral 7B/8x7B)No (API only)

Frequently Asked Questions

Does Mistral have a free tier?

Yes. Mistral's la Plateforme offers a free tier with rate-limited API access for testing. The free tier covers all models but is not suitable for production. Commercial use requires a paid plan with per-token billing.

Can I self-host Mistral models?

Yes — Mistral 7B, Mistral 8x7B (Mixtral), and Mistral 8x22B are open-weight models you can download and self-host via Ollama, vLLM, or llama.cpp. Self-hosting eliminates per-token costs (you pay only for GPU time). Mistral Large is API-only.

What is the Mistral API rate limit?

Free tier: 1 request/second. Paid tiers vary by plan — most paid tiers start at 60 RPM and scale up with spend. Contact Mistral for enterprise volume limits.

Calculate your exact Mistral vs Claude vs GPT-4o cost for your specific token volume.

Open Pricing Calculator →