Mistral AI September 2024 Release

Mistral AI has launched a free tier for its serverless platform, reduced pricing for its entire model family, and released Mistral Small v24.09. These updates aim to lower the barrier to entry for developers and provide more cost-efficient, enterprise-grade options for specific AI tasks.

Free Tier for la Plateforme

la Plateforme, Mistral AI's serverless platform for tuning and building with models as API endpoints, now includes a free tier. This tier allows developers to perform experimentation, evaluation, and prototyping at no cost.

Users can transition from the free tier to a commercial tier to access higher rate limits and full data isolation, including a free zero-retention option. Deployment options remain flexible, allowing users to utilize cloud partners such as Azure, AWS, and GCP, or deploy via their own tenant.

Model Pricing Reductions

Mistral AI has reduced prices across its entire model family to improve the cost-efficiency of its endpoints. These changes apply to la Plateforme and cloud partner offerings including Azure AI Studio, Amazon Bedrock, and Google Vertex AI.

Model New Price (Input/Output per M tokens) Old Price (Input/Output per M tokens) Price Drop
Mistral Nemo $0.15 / $0.15 $0.3 / $0.3 50%
Pixtral 12B $0.15 / $0.15 N/A N/A
Mistral Small $0.2 / $0.6 $1 / $3 80%
Codestral $0.2 / $0.6 $1 / $3 80%
Mistral Large $2 / $6 $3 / $9 33%

According to Mistral AI, these updates make Mistral Large 2 the most cost-efficient frontier model available.

Mistral Small v24.09

Mistral Small v24.09 is an enterprise-grade small model with 22 billion parameters, replacing Mistral Small v24.02. It is designed as a cost-effective mid-point between Mistral NeMo 12B and Mistral Large 2.

Released under the Mistral Research License (MRL), the model is optimized for tasks such as translation, summarization, and sentiment analysis that do not require a general-purpose model. Mistral AI states that v24.09 delivers significant improvements in reasoning capabilities, human alignment, and code over its predecessor.

Pixtral 12B on le Chat

Pixtral 12B, a vision-capable model, is now available for free on le Chat. This model allows users to scan, analyze, search, and caption images and knowledge files.

Pixtral 12B is the first open-source model to support images of any size without degrading text-based performance. Because it is available under the Apache 2.0 license, users can deploy the model in their own environments to maintain control over sensitive or proprietary information without relying on third-party providers.

Sources

Related

  • Dispatch
    Mistral AI Devstral 2 and Mistral Vibe CLI Release

    Mistral AI has released Devstral 2 (123B) and Devstral Small 2 (24B), open-weight coding models that achieve state-of-the-art performance on SWE-bench Verified with high cost-efficiency, alongside the open-source Mistral Vibe CLI for autonomous code automation.

  • Dispatch
    Mistral AI Devstral Models Release

    Mistral AI has released Devstral Medium and Devstral Small 1.1, providing state-of-the-art agentic coding capabilities with a focus on generalization across different prompts and agentic scaffolds.

  • Dispatch
    Mistral Small 3.1 Release Notes

    Mistral AI has released Mistral Small 3.1, an Apache 2.0 licensed model featuring multimodal understanding, a 128k token context window, and performance that surpasses GPT-4o Mini and Gemma 3.

  • Dispatch
    Mistral AI La Plateforme Release

    Mistral AI has launched La Plateforme, a beta API service providing access to three generative chat endpoints and one embedding endpoint with varying performance and price tradeoffs.

  • Dispatch
    Mistral Medium 3 Release Notes

    Mistral AI has released Mistral Medium 3, a frontier-class model that balances state-of-the-art performance with 8x lower costs and flexible enterprise deployment options.