Mistral Small 3 ( 2501 ) sets a new benchmark in the "small" Large Language Models category below 70B, boasting 24B parameters and achieving state-of-the-art capabilities comparable to larger models!
Fine-tuningDocs | Mistral Small 24B Instruct 2501 can be customized with your data to improve responses. Fireworks uses LoRA to efficiently train and deploy your personalized model |
On-demand DeploymentDocs | On-demand deployments allow you to use Mistral Small 24B Instruct 2501 on dedicated GPUs with Fireworks' high-performance serving stack with high reliability and no rate limits. |
Mistral Small 24B Instruct 2501 is an instruction-tuned version of the base Mistral Small 24B model, developed by Mistral AI. It is designed as a high-performance, "small" LLM (under 70B parameters) that competes with much larger models. It supports multilingual tasks and is well-suited for chat, reasoning, and structured output generation.
Its performance is validated across generalist, reasoning, and coding benchmarks.
The model supports a context window of 32,768 tokens.
The full 32.8K token context window is available on Fireworks' on-demand deployments with no rate limits.
The model has 23.6 billion parameters.
Yes. Fireworks supports LoRA-based fine-tuning through its RFT infrastructure.
Fireworks charges based on combined input + output token usage.
The model is released under the Apache 2.0 license, allowing unrestricted commercial use.
