AGENTKERNEL
  1. Home
  2. All models
  3. Mixtral 8x22B Instruct

Mixtral 8x22B Instruct: price, performance and context

Open Mixtral 8x22B Instruct by Mistral costs $2.00 per 1M input and $6.00 per 1M output tokens, handles up to 65.5K tokens in one request (≈ 200 pages of text) and scores 5.7 points on the AgentKernel performance index, rank 204 of 212.

Privacy: zero data retention with one provider, including in the EU. Availability over the last 24 h: 96.2%.

Updated Oct 09, 2026, 16:49 UTC · mistralai/mixtral-8x22b-instruct

Open in the interactive comparison

Key facts

Licence
Open weights: free to download and self-host
Input / 1M tokens
$2.00
Output / 1M tokens
$6.00
Context window
65.5K tokens
Max. output
52.4K tokens
Performance index
5.7 points · rank 204 of 212
Coding index
n/a
Agentic index
n/a
Cost at your usage
$3.50 / month
Knowledge cutoff
Jan 2024
In the catalog since
Apr 17, 2024
Capabilities
Tool Calling, Files

What do you get for $10.00?

Rules of thumb from the cost calculator, base rates without caching or taxes.

TaskCost per taskFor $10.00
Chat repliesA question with a short history, a detailed answer$0.0042,800
Email draftsBullet points in, a finished email out$0.0033,400
Document summariesRead 10 pages, write a one-page summary$0.011,000
Code reviewsReview and comment on 500 lines of code$0.017590
Book analysesAnalyse a 300-page book in one go–does not fit the context

Cheaper with similar performance

ModelLicencePerformanceYour cost / month
Ling 3.1 FlashinclusionAIProprietary41.1$0.00
Ling 3.0 Flash VLinclusionAIOpen24.6$0.036
Ling 3.0 FlashinclusionAIOpen20.1$0.037

Stronger models

ModelLicencePerformanceYour cost / month
Mistral LargeMistralProprietary5.8$3.50
Phi 4MicrosoftOpen5.9$0.105
Nova Micro 1.0AmazonProprietary5.9$0.07