How much does Mistral: Mistral Small 3 cost?

Mistral: Mistral Small 3 costs $0.05 per 1M input tokens and $0.08 per 1M output tokens.

What is the context length of Mistral: Mistral Small 3?

Mistral: Mistral Small 3 has a 28,000 token context window.

Does Mistral: Mistral Small 3 support tool calling and structured outputs?

Yes. Mistral: Mistral Small 3 supports structured outputs.

What inputs and outputs does Mistral: Mistral Small 3 support?

Mistral: Mistral Small 3 accepts text as input and returns text.

What other models does MistralAI have?

Mistral: Mistral Nemo is another model from MistralAI.

When was Mistral: Mistral Small 3 released?

Mistral: Mistral Small 3 was released on Jan 30, 2025.

Mistral Small 3

MistralAIReleased Jan 30, 2025mistralai/Mistral-Small-24B-Instruct-2501

Cost Compare

A 24B parameter model optimized for speed and efficiency while rivaling models 3x its size like Llama 3.3 70B. Achieves 81% accuracy on MMLU benchmarks with impressive 150 tokens/second throughput on consumer hardware. Released under Apache 2.0 license, it serves as an excellent base for developing advanced reasoning capabilities.

Context

28K

Max output

16K

Input /1M

$0.05

Output /1M

$0.08

Blended /1M

$0.06

AcceptsTextProducesTextTokenizer Mistral

Providers

FastRouter routes your requests to this provider. You can also bring your own key.

ProviderInput /1MOutput /1MContextMax outputLatencyThroughput

DeepInfrafp8

deepinfra

Input /1M$0.05

Output /1M$0.08

Context28K

Max output16K

Latency—

Throughput—

Supported parameters

Request parameters you can send with this model.

Core (Sampling & basic generation)

max_tokens: Maximum tokens to generate.
min_p: Minimum probability threshold for sampling.
seed: Seed for deterministic sampling.
stop: Sequences where generation stops.
temperature: Controls randomness of output.
top_k: Limits sampling to top K tokens.
top_p: Nucleus sampling threshold.

Penalties (Token Penalties)

frequency_penalty: Penalizes frequent tokens.
presence_penalty: Penalizes tokens already present.
repetition_penalty: Penalizes repeated tokens.

Reasoning (Reasoning Controls)

include_reasoning: Include reasoning trace in the response.
reasoning: Provider-specific reasoning configuration.

Output & Format (Response Formatting)

logit_bias: Bias values for specific token IDs.
logprobs: Return log probabilities of output tokens.
response_format: Structured output format configuration.

Performance

Median throughput and time to first token per provider over the last week.

Throughput

Latency

Make your first API call

OpenAI-compatible. Point your SDK at FastRouter, or call the API directly.

Full API docs

Replace <FASTROUTER_API_KEY> with your key · Get your key →Install

OpenAI SDKHTTP

from openai import OpenAI

client = OpenAI(

base_url="https://api.fastrouter.ai/api/v1", # FastRouter base URL

api_key= "<FASTROUTER_API_KEY>", # Replace with your FastRouter API key

)

completion = client.chat.completions.create(

model="mistralai/Mistral-Small-24B-Instruct-2501", # Replace with your model ID

messages=[

{ "role": "user", "content": "What is the meaning of life?" }

]

)

print(completion.choices[0].message.content)

Frequently asked questions

More models from MistralAI

Mistral Nemomistralai/Mistral-Nemo-Instruct-2407

A streamlined 12B parameter model developed in collaboration with NVIDIA featuring 128K context window support. Offers multilingual capabilities and integrated function calling, making it well-suited for deployment on NVIDIA hardware with optimized performance characteristics for real-time applications.

Model scores sourced from ArtificialAnalysis