Ministral 3 8B 2512

Provider: Mistral AI EU Official API price

Input price
$0.15
/ 1M tokens
Output price
$0.15
/ 1M tokens
Cache price
$0.015
/ 1M tokens (read)
Context
262K tokens
Max output 210K tokens

Overview

Ministral 3 8B 2512 from Mistral AI (EU) is a mid-tier model designed for long-context and multimodal workloads. It supports vision capabilities, function calling, and JSON mode, making it versatile for complex applications requiring structured output and tool interaction.

Strengths

  • Boasts an exceptionally large 262,144 token context window for extensive document processing.
  • Features vision capabilities for understanding and processing image inputs.
  • Includes function-calling and JSON mode, enabling structured outputs and integration with tools.
  • Offers highly competitive symmetrical pricing for both input and output tokens at $0.15/M.
  • Provides an extremely low cache read price of $0.015/M, beneficial for repeated access tasks.

Best use cases

  • Advanced long-document analysis, summarization, and question-answering over large texts.
  • Multimodal applications requiring visual understanding alongside textual processing.
  • Developing agentic workflows that leverage function calling for external tool integration.
  • Generating structured data or API responses consistently using JSON mode.
  • Applications with frequent, recurring data access where the low cache read price is advantageous.

Pricing position

Ministral 3 8B 2512 is priced at $0.15/M for both input and output, positioning it on the higher side for input compared to Mistral Nemo ($0.02/M) and Mistral Small 3 ($0.05/M) within its family. However, its $0.15/M output price is significantly more economical than Mistral Small 3.2 ($0.20/M) and Voxtral Small ($0.30/M), and substantially cheaper than peer models like Llama 4 Maverick ($0.6/M), GPT-4o-mini ($0.6/M), and MiniMax M2.5 ($1.15/M).

Caveats

The maximum output token limit for Ministral 3 8B 2512 is not specified, which could be a consideration for generating very lengthy responses. This model does not offer a free tier for initial testing or low-volume usage. For tasks not requiring its large context or specialized capabilities, other Mistral AI models like Nemo or Small 3 might offer a more cost-effective solution.

Notes drafted by an LLM from the figures on this page (2026-05-21) and rewritten when the price changes.

Key facts

Price basis
Official API price
Max output
210K tokens
Listed on OpenRouter
2025-12-02
Tracked since
2026-05-08
Last updated
2026-11-10

Features

Image inputFunction callingJSON output

Monthly cost estimate

Scale Input (M) Output (M) Monthly
0.1M (hobby) 0.1 0.05 $0.02
1M (small app) 1 0.5 $0.22
10M (mid scale) 10 5 $2.25
100M (large) 100 50 $22.50

※ Cache and batch discounts not factored in. Detailed calculator

Official resources

Other models from Mistral AI

Similar-priced models

Source: OpenRouter API (mistralai/ministral-8b-2512)