impulse

Mistral Small 3.1 24B Instruct 2503

Open weights

Mistral Small 3.1 24B Instruct is an upgraded variant of Mistral Small 3 (2501), featuring 24 billion parameters with advanced multimodal capabilities.

API id mistral-small-3-1-24b-instruct-2503Released 17 Mar 2025 (Mar 2025)
Input / 1M
$0.106
at IBM watsonx.ai
Output / 1M
$0.318
at IBM watsonx.ai
Context window
131K
tokens
Max output
16K
tokens per reply
Hosts
1
1 listings
Released
Mar 2025
17 Mar 2025

Pricing

List prices at IBM watsonx.ai, in US dollars per million tokens.

TierInput / 1MOutput / 1M
Standard$0.106$0.318

What it costs a month

At IBM watsonx.ai list prices, 30 days.

  • Chat assistant5,000 requests a day, 2K in / 0.5K out$55.65
  • RAG over documents2,000 requests a day, 8K in / 0.6K out$62.33
  • Coding agent1,000 requests a day, 40K in / 2K out$146
  • Overnight batch20,000 requests a day, 4K in / 0.4K out$331
Use your own numbers

Where to run it

1 host serve Mistral Small 3.1 24B Instruct 2503. The comparison is on blended price, 3 parts input to 1 part output.

HostIn / out per 1M
mistral-small-3-1-24b-instruct-2503
$0.106 / $0.318

Capabilities

Core

  • Streaming
  • Function calling
  • JSON mode
  • Structured outputs

Multimodal

  • Vision
  • Image input
  • Audio
  • Image generation
  • Video

Advanced

  • Reasoning
  • Tool calling
  • MCP
  • Prompt caching
  • Batch API
  • Embeddings
Takes in
Text
Produces
Text
Features
Parallel tool calls
Open weights
License: Apache-2.0
Runs in about 57.6 GB of GPU memory

Cheaper alternatives

Current models that cost less and keep most of what this one can do: tools, vision, reasoning, structured output, caching and enough context.

Same model, cheaper host

No host sells it for less than IBM watsonx.ai.

Smaller Mistral AI models

  • Devstral Small
    Mistral AI · on Mistral AI
    $0.10 / $0.30
    Has every key capability
    6% less

Other frontier labs

No cheaper match from another frontier lab.

Open-weight picks

  • Ministral 8B 2512
    Mistral AI · on Mistral AI
    $0.15 / $0.15
    Has every key capability
    6% less
  • DeepSeek V4.1 Flash
    DeepSeek · on Sail
    $0.09 / $0.30
    Has every key capability
    10% less
  • MiMo V2.6 Flash
    Xiaomi · on GMI Cloud
    $0.112 / $0.224
    Has every key capability
    12% less
  • Qwen3.6 35B A3B
    Alibaba Qwen · on Sail
    $0.05 / $0.40
    Has every key capability
    14% less

Call it

Using IBM watsonx.ai's API.

Python
from litellm import completion

response = completion(
    model="watsonx/mistral-small-3-1-24b-instruct-2503",
    messages=[{"role": "user", "content": "Hello"}],
)
print(response.choices[0].message.content)

Quick answers

How much does Mistral Small 3.1 24B Instruct 2503 cost?

On IBM watsonx.ai, Mistral Small 3.1 24B Instruct 2503 costs $0.106 per million input tokens and $0.318 per million output tokens.

Which APIs offer Mistral Small 3.1 24B Instruct 2503?

Mistral Small 3.1 24B Instruct 2503 runs on 1 API: IBM watsonx.ai.

What is Mistral Small 3.1 24B Instruct 2503's context window?

131K tokens, with up to 16K tokens in one reply.

Is Mistral Small 3.1 24B Instruct 2503 being retired?

No retirement date is scheduled for it on any API we track.

What are cheaper alternatives to Mistral Small 3.1 24B Instruct 2503?

Models that cost less and keep most of what it can do: Devstral Small on Mistral AI (6% less) and Ministral 8B 2512 on Mistral AI (6% less). Savings are on a mix of 3 input tokens to 1 output token.

When was Mistral Small 3.1 24B Instruct 2503 released?

Mistral AI released it on 17 Mar 2025.

What is Mistral Small 3.1 24B Instruct 2503's API model id?

Use mistral-small-3-1-24b-instruct-2503 with IBM watsonx.ai.

Sources

Raw catalog entry for mistral-small-3-1-24b-instruct-2503 on IBM watsonx.ai
{
  "max_input_tokens": 131072,
  "max_output_tokens": 16384,
  "input_cost_per_token": 1.06e-7,
  "output_cost_per_token": 3.18e-7,
  "litellm_provider": "watsonx",
  "mode": "chat",
  "supports_function_calling": true,
  "supports_parallel_function_calling": true,
  "supports_vision": false,
  "source": "https://dataplatform.cloud.ibm.com/docs/content/wsj/analyze-data/fm-models.html?context=wx"
}