Mistral · Model
Codestral 2508 pricing, context window and specs
Codestral 2508 costs $0.30 per million input tokens and $0.90 per million output tokens, and reads up to 128,000 tokens per request. By blended price it is number 6 of Mistral’s 8 current models here, counting from the cheapest.
- Input / 1M
- $0.30
- Output / 1M
- $0.90
- Cached / 1M
- $0.03
- Context
- 128,000
- Max output
- Not published
- Released
- 2025-08-01
Estimate your cost with Codestral 2508Compare all Mistral modelsCount tokens for Codestral 2508
Pricing
Codestral 2508 API pricing
| Price per 1M tokens | Input | Output |
|---|---|---|
| Standard | $0.30 | $0.90 |
| Cached input (read) | $0.03 | – |
| Batch API | $0.15 | $0.45 |
Output costs 3.0× the input price, so long replies drive the bill more than long prompts. If your requests share a long fixed prefix, such as a system prompt or tool definitions, caching cuts that part to 10% of the normal input price.
Examples
What Codestral 2508 costs in practice
| Workload | Tokens in / out | Requests/day | Per request | Per month |
|---|---|---|---|---|
| Support chatbot50% cached | 1,500 / 400 | 1,000 | $0.000608 | $18.48 |
| RAG app20% cached | 6,000 / 500 | 500 | $0.00193 | $29.29 |
| Coding agent80% cached | 40,000 / 2,000 | 200 | $0.00516 | $31.39 |
| Document summariserbatch | 8,000 / 600 | 300 | $0.00147 | $13.41 |
Change any number in the cost calculator.
Context
Context window
Codestral 2508 accepts up to 128,000 tokens per request, roughly 96,000 English words or 192 pages. That budget is shared between your prompt and the reply, and no separate cap on the reply is published.
Filling the whole window with one prompt costs about $0.0384 at list price, so for repeated long-document work, retrieval (sending only the relevant parts) or prompt caching is usually much cheaper.
Features
Capabilities
- Image input
- No
- Tool / function calling
- Yes
- Reasoning mode
- No
- Prompt caching
- Yes
- Structured output
- Yes
- Open weights
- No
- Accepts
- text, file
- Knowledge cutoff
- 2025-03-31
As declared by the provider’s API. It shows a feature exists, not how well it works.
API
Codestral 2508 model ID
- OpenRouter
mistralai/codestral-2508
Use the OpenRouter ID when calling it through OpenRouter; Mistral’s own API may name it differently. See the model ID reference for the providers we document.
Mistral
Where Codestral 2508 sits in Mistral’s line-up
Ranked by blended price (three parts input to one part output, $0.45 per 1M for Codestral 2508), it is number 6 of Mistral’s 8 current models here, counting from the cheapest. One step down, Mistral Small 4 costs $0.2625 blended (42% less). Unlike Codestral 2508, it accepts images, doesn’t accept files and reads up to 262,144 tokens (128,000 here). One step up, Mistral Large 4 costs $1.0325 (2.3× as much). Unlike Codestral 2508, it accepts images, doesn’t accept files and reads up to 1,048,576 tokens (128,000 here). See every Mistral model and price
| Model | Input / 1M | Output / 1M | Context | Released |
|---|---|---|---|---|
| Ministral 3 3B 2512 | $0.10 | $0.10 | 131,072 | 2025-12-02 |
| Ministral 3 8B 2512 | $0.15 | $0.15 | 262,144 | 2025-12-02 |
| Voxtral Small 24B 2507 | $0.10 | $0.30 | 32,768 | 2025-10-30 |
| Ministral 3 14B 2512 | $0.20 | $0.20 | 262,144 | 2025-12-02 |
| Mistral Small 4 | $0.15 | $0.60 | 262,144 | 2026-03-16 |
| Codestral 2508 (this page) | $0.30 | $0.90 | 128,000 | 2025-08-01 |
| Mistral Large 4 | $0.68 | $2.09 | 1,048,576 | 2026-10-06 |
| Mistral Medium 3.5 | $1.50 | $7.50 | 262,144 | 2026-04-30 |
Mistral’s current models with a page here, cheapest first. Older and superseded models are listed on the provider page and in the comparison table.
Save
Cheaper alternatives
- Ministral 3 3B 2512Mistral78% cheaper
- Nemotron 3.5 LightningNVIDIA77% cheaper
- Gemma 4 26B A4BGoogle76% cheaper
- Qwen3.5-9BQwen75% cheaper
- Seed 1.6 FlashByteDance Seed71% cheaper
Current models from major providers, one per provider, that keep tool calling, a usable context window and output limit. Compared by blended price.
Compare
Similarly priced models
- GLM 4.6VZ.ai$0.30 / $0.90
- DeepSeek V3.2DeepSeek$0.259 / $0.80
- Muse Glimmer 30BMeta$0.30 / $1.20
- MiniMax M3MiniMax$0.30 / $1.20
- Qwen3 VL 8B InstructQwen$0.25 / $0.75
Current models from other major providers, closest blended price.
FAQ
Codestral 2508 questions
How much does Codestral 2508 cost?
Codestral 2508 costs $0.30 per million input tokens and $0.90 per million output tokens, with cached input at $0.03. A typical chatbot reply (1,500 tokens in, 400 out) costs about $0.00081. Use the LLM cost calculator for your own workload.
What is Codestral 2508’s context window?
128,000 tokens, about 96,000 English words or 192 printed pages, shared between your prompt and the reply. No separate limit on the reply is published, so the window is the limit.
Does Codestral 2508 support prompt caching and batch requests?
Yes, cached input is billed at $0.03 per million tokens (10% of the normal input price). A batch API is available at $0.15 input and $0.45 output per million tokens.
What are cheaper alternatives to Codestral 2508?
With the same essentials (tool calling), the cheapest options are Ministral 3 3B 2512 (78% cheaper), Nemotron 3.5 Lightning (77% cheaper), Gemma 4 26B A4B (76% cheaper). Cheaper doesn’t mean equivalent, so test them on your own prompts.
What is the API model ID for Codestral 2508?
On OpenRouter it is mistralai/codestral-2508. Mistral’s own API may use a different name, so check its model list before you deploy. Pass the ID as the model value in each request.
Is Codestral 2508 open source?
No, Codestral 2508 is only available through Mistral’s API and partners; its weights aren’t published.