Codestral 2508
Mistral's cutting-edge language model for coding released end of July 2025. Codestral specializes in low-latency, high-frequency tasks such as fill-in-the-middle (FIM), code correction and test generation. [Blog Post](https://mistral.ai/news/codestral-25-08)
codestral-2508Released 1 Aug 2025 (Aug 2025)Pricing
List prices at Mistral AI, in US dollars per million tokens.
| Tier | Input / 1M | Output / 1M | |
|---|---|---|---|
| Standard | $0.30 | $0.90 | |
| Cached input 90% off input | $0.03 | 90% off input |
What it costs a month
At Mistral AI list prices, 30 days.
- Chat assistant5,000 requests a day, 2K in / 0.5K out, cached$133
- RAG over documents2,000 requests a day, 8K in / 0.6K out, cached$112
- Coding agent1,000 requests a day, 40K in / 2K out, cached$187
- Overnight batch20,000 requests a day, 4K in / 0.4K out$936
Where to run it
1 host serve Codestral 2508. The comparison is on blended price, 3 parts input to 1 part output.
| Host | In / out per 1M | Model id | |
|---|---|---|---|
Mistral AILab API codestral-2508 | $0.30 / $0.90 | codestral-2508 |
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
- Assistant prefill
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 Mistral AI.
Smaller Mistral AI models
- 42% less
- 42% less
- 47% less
Other frontier labs
- 56% less
- 56% less
- 42% less
Open-weight picks
- 3% less
- 18% less
- 22% less
- 42% less
Model line
Newer and older versions, at each lab's own price.
- Aug 2025Codestral 2508Latest$0.30 / $0.90
- UndatedCodestral 2501$0.20 / $0.60
- UndatedCodestral 2405Not listed
- UndatedCodestral 2$0.30 / $0.90
Call it
Using Mistral AI's API.
from litellm import completion
response = completion(
model="mistral/codestral-2508",
messages=[{"role": "user", "content": "Hello"}],
)
print(response.choices[0].message.content)Quick answers
How much does Codestral 2508 cost?
On Mistral AI, Codestral 2508 costs $0.30 per million input tokens and $0.90 per million output tokens. Cached input costs $0.03 per million tokens.
Which APIs offer Codestral 2508?
Codestral 2508 runs on 1 API: Mistral AI.
What is Codestral 2508's context window?
128K tokens, with up to 128K tokens in one reply.
Is Codestral 2508 being retired?
No retirement date is scheduled for it on any API we track.
Is there a newer version of Codestral 2508?
No. Codestral 2508 is the newest model in its line.
What are cheaper alternatives to Codestral 2508?
Models that cost less and keep most of what it can do: Magistral Small on Mistral AI (42% less), Claude Haiku 5.5 on Anthropic (56% less), GPT-6 Luna on OpenAI (56% less) and MiMo V2.6 Pro on GMI Cloud (3% less). Savings are on a mix of 3 input tokens to 1 output token.
When was Codestral 2508 released?
Mistral AI released it on 1 Aug 2025.
What is Codestral 2508's API model id?
Use codestral-2508 with Mistral AI.
Sources
Raw catalog entry for codestral-2508 on Mistral AI
{
"cache_read_input_token_cost": 3e-8,
"input_cost_per_token": 3e-7,
"litellm_provider": "mistral",
"max_input_tokens": 128000,
"max_output_tokens": 128000,
"mode": "chat",
"output_cost_per_token": 9e-7,
"source": "https://docs.mistral.ai/models/model-cards/codestral-25-08",
"supports_assistant_prefill": true,
"supports_function_calling": true,
"supports_response_schema": true,
"supports_tool_choice": true
}