Mistral pricing: $0 to $24.99 a month, API from $0.10

In brief

Mistral pricing is $0 on Vibe Free, $14.99 a month on Pro and $24.99 per user on Team, as of November 2026, and Mistral lists no yearly plan. The API bills per token: Mistral Large 4 costs $0.70 per million input and $2.10 per million output, so a standard answer costs $0.00175, a quarter of Claude Sonnet 5.5.

How much does Mistral cost on each plan?

Mistral costs $0 to $24.99 a month for the Vibe chat app, and $0.10 to $2.10 per million tokens on the API. We read every figure below from Mistral's pricing page and help center on 14 November 2026. Vibe is what Mistral renamed Le Chat on 28 May 2026. Prices exclude tax:

Four frosted glass blocks of rising height on black plinths, an amber one on the shortest and a clear one on the tallest, a chrome ruler leaning against them and a small terracotta cube set apart, in a charcoal studio.
Plan, read 14 November 2026PriceWhat Mistral lists
Vibe Free$0Limited messages and web searches, limited coding sessions, basic image generation and editing, 100+ connectors
Vibe Pro$14.99 a monthMore messages and web searches, faster replies, coding in the CLI, IDE or web, spreadsheets and slides, better image generation, chat and email support
Education (student Pro)$5.99 a month, per the FAQMistral Pro for verified students, with an institutional email
Vibe Team$24.99 per user a monthHigher limits, faster replies, up to 30 GB of storage per user, domain verification, data export
EnterpriseThrough salesCustom models, agents and workflows, audit logs, SAML SSO, white label

Mistral lists no yearly price for any of the plans, so there is no yearly discount that you could chase. The pricing FAQ says "Verified students can get Mistral Pro for $5.99 / month (normally $14.99)." Mistral's student help article says you must verify an institutional email address, and it gives no price, length or once-only rule. The saving is $9 a month. The Pro card shows $14.99 with an "I'm a student" button beside it, so check the price at checkout.

What does the Mistral API cost per token?

The Mistral API costs $0.70 per million input tokens and $2.10 per million output tokens on Mistral Large 4, its flagship, as of November 2026. Smaller models cost far less. Mistral's API price list shows these generalist rates, read 14 November 2026:

ModelInput per 1M tokensCached inputOutput per 1M tokens
Mistral Large 4$0.70$0.07$2.10
Mistral Small 4$0.15$0.015$0.60
Ministral 3 14B$0.20$0.02$0.20
Ministral 3 8B$0.15$0.015$0.15
Ministral 3 3B$0.10$0.01$0.10
Codestral$0.30$0.03$0.90
Mistral Embed$0.10$0.01none

The same list also prices OCR 4.1 at $4 for every 1,000 pages and OCR 3 at $2. If you need speech, Voxtral Mini Transcribe 3 costs $0.0035 a minute, and Voxtral TTS 3 costs $13 for each million characters it speaks. Mistral's launch post for Large 4 is dated 6 October 2026 and calls it a public preview, with the API open on Mistral Studio. It lists no price. The pricing page FAQ adds that batch processing cuts the price by 50% and that cached input tokens cut input cost by up to 90%. That FAQ also says "Mistral Large costs $0.5 /M tokens in and $1.5 /M tokens out," which doesn't match the $0.70 and $2.10 on the price list, so we used the price list.

That list isn't all Mistral. It also resells Z.ai GLM 5.3 at $1.40 in and $4.40 out, and DeepSeek V4.1 Flash at $0.30 and $1.20. Those aren't Mistral's own models, so we left them out of the table. They do show that the price list doubles as a catalog of other labs' models.

Is Mistral cheaper than Claude?

Mistral Large 4 is cheaper than Claude Sonnet 5.5 and Opus 5.5, but not cheaper than Claude Haiku 5.5. Take a standard answer of 1,000 tokens in and 500 out. On Large 4 it costs $0.70 per million times 0.001 plus $2.10 per million times 0.0005, which is $0.00175. On Sonnet 5.5 at $2 and $10, it costs $0.007.

Large 4 costs a quarter of Claude Sonnet 5.5 per answer, and five times Claude Haiku 5.5
0.01 cent0.1 cent1 centMinistral 3 3BClaude Haiku 5.5Mistral Small 4Mistral Large 4Claude Sonnet 5.5Claude Opus 5.5

Here are the same figures for 1,000 answers, with each vendor's price read from its own page on 14 November 2026. The Claude rates come from Anthropic's pricing docs, for prompts under 100,000 tokens:

ModelInput and output per 1M tokensCost of 1,000 answers
Ministral 3 3B$0.10 and $0.10$0.15
Claude Haiku 5.5$0.10 and $0.50$0.35
Mistral Small 4$0.15 and $0.60$0.45
Mistral Large 4$0.70 and $2.10$1.75
Claude Sonnet 5.5$2 and $10$7.00
Claude Opus 5.5$4 and $20$14.00

Cheap per token isn't the same as good per task. Price is only half of the story, and the place to check what Mistral claims about quality is its own Large 4 post. The Le Chat vs ChatGPT page covers how Large 4 scores against OpenAI's models, and Claude API pricing has the full Anthropic list with caching and batch rates.

What does Vibe Free include, and where does it stop?

Vibe Free includes limited messages, web searches and coding sessions, plus basic image generation, and Mistral publishes no number for those caps. Its pricing page uses the word "limited" and nothing else for the Free card. The feature matrix prints multiples in the paid columns: messages "Up to 6x free," web searches "Up to 5x free" and image generation "Up to 40x free." They give no absolute number to multiply. It does print a few hard numbers: Free allows up to 5 scheduled tasks against unlimited on Pro, and the FAQ says Flash answers are 150 a day on Pro and 200 a day on Team.

For the API side, the docs are blunt. The usage page says "Free mode includes a monthly usage budget for trying Vibe and Studio." It says free mode doesn't include API access, and that creating API keys needs billing activation and pay-as-you-go. The help center adds that free mode "has the lowest limits, intended for evaluation and prototyping," with no figures.

If a Free cap you can't measure is the problem, Whizi's free plan is the other thing to try. It gives 10 messages a day on Whizi AI, Auto and Claude Haiku 5.5, 2 pictures a month and 5 photo uploads a day, with no card. Whizi Pro carries Claude, GPT and Gemini, and its first week is $1.99. Whizi doesn't run Mistral's Vibe app, so it replaces Vibe only if you want other labs' models in one chat.

Is Vibe Pro worth $14.99 a month?

Vibe Pro is worth $14.99 if you use its API credits, because the credit alone roughly matches the price. Mistral's pricing page prints two feature lists with API credits: "$25.5 /mo in API credits" under the $14.99 Pro price, and "$15 /mo in API credits" in a second list that carries no price. Mistral doesn't say which applies to which reader, so we use the lower $15 here. Check it before you buy.

At $15, the credit covers about 8,570 standard answers on Large 4 at $0.00175 each. On Claude Sonnet 5.5 the same $15 would cover about 2,140. That's a derived figure, not a Mistral claim, and it assumes you spend the credit on one model. If you never touch the API, Pro buys more messages, faster replies and coding in the CLI and IDE. Mistral publishes no cap for those, so you're paying for headroom that you can't size in advance, and the only way to learn whether it's enough is to try a month. That's a real cost.

Team at $24.99 per user is $9.99 more than Pro for higher limits, 30 GB of storage per user and domain verification. Ten users cost $249.90 a month. Choose it for the admin features, since Mistral lists no API credit for Team that we could find, and it would not make sense to buy it for the credit anyway.

What are the Mistral API rate-limit tiers?

Mistral's API limits depend on your usage tier, and tiers rise by what you've been billed, not by what you prepay. Its help center says "Rate limits depend on your plan and usage tier." Pay-as-you-go unlocks Tier 1 and above. Tiers 2 to 4 start at cumulative billed totals above €20, €100 and €500, or the dollar equivalents.

A 429 error means you passed one of three limits: requests per second, tokens per minute or tokens per month, and limits are set per model. Free mode has the lowest of them, and prepaid credit buys usage but doesn't change the tier that you are on. Once you've been billed more than €2,000, you can contact support with the model you use, the requests per second that you want and your estimated tokens per minute and per month. The help center says that "With pay-as-you-go, tiers upgrade automatically based on your cumulative billed amount." So if you're waiting on a higher limit, topping up credit won't lift it by itself, because only the money you actually spend on usage counts toward the next tier. So the way to get a higher tier is to use the service and pay for it.

Which Mistral plan should you pick?

Pick Vibe Free to test, Pro at $14.99 for daily coding and chat, and the API for anything inside your own app. Students should take the $5.99 plan, because it is the same product for $9 less.

Pick the API over Pro if you run more than a few thousand answers a month, since Large 4 costs only $1.75 per 1,000 standard answers and you pay for exactly what you use. Pick Team if your company needs shared storage and domain checks. Pick Enterprise only if you need SSO, audit logs or private deployment, which Mistral prices through sales. If you also pay for another chat app, Mistral AI explained covers what Vibe does that the others don't.

Workflow checklist
  • Start on Vibe Free at $0 and pay $14.99 for Pro only when you hit the cap
  • Students: verify an institutional email for the $5.99 plan, and check the price at checkout since the card shows $14.99
  • Price your API work on Large 4 at $1.75 per 1,000 standard answers before you pick a model
  • Compare against Claude Haiku 5.5 at $0.35 per 1,000 answers if cost matters more than the Mistral name
  • Remember that Free mode has no API access, so API work starts with billing activation
  • Spend a little through pay-as-you-go to reach Tier 2 at €20, since prepaid credit alone won't raise limits
Common questions

Frequently asked questions

Is Mistral free?

Partly. Vibe Free costs $0 and includes limited messages, web searches, coding sessions and basic image generation, as of November 2026, but Mistral publishes no number for those caps. Its docs say free mode has no API access. For what else Mistral sells, see Mistral AI explained.

How much is Mistral Pro?

Mistral Pro costs $14.99 a month, excluding tax, as of November 2026, with no yearly price listed. Verified students pay $5.99 a month, and Team costs $24.99 per user a month. See Le Chat vs ChatGPT for how that compares with ChatGPT Plus at $20.

How much does the Mistral API cost?

Mistral Large 4 costs $0.70 per million input tokens and $2.10 per million output tokens, as of November 2026. Mistral Small 4 costs $0.15 and $0.60, and Ministral 3 3B costs $0.10 on both sides. Batch processing cuts the price by 50%, per Mistral's FAQ. Claude API pricing lists Anthropic's rates.

Is Mistral cheaper than ChatGPT?

On the chat app, yes: Vibe Pro costs $14.99 a month against $20 for ChatGPT Plus. On the API, Large 4 costs $1.75 per 1,000 standard answers against $7 for Claude Sonnet 5.5, and Claude Haiku 5.5 beats it at $0.35. Qwen pricing shows another low-cost lab.