LMFinder
C

Codestral 2508

Reviewed 2026-06-27
Mistralmistralai/codestral-2508
Context
256K
Price
$0.3/$0.9
Tooling
3/5
Deploy
Cloud API

Best for

  • Writing and editing code
  • Long documents (~256K context)
  • Cost-sensitive, high-volume workloads

Skip if

  • You need a specialized, domain-tuned model

Capability scores

Tool calling3/5
Multi-step reasoning3/5

Specifications

Context window
256K tokens
VRAM
N/A (cloud-hosted)
License
mistral-ai-non-production
Pricing
$0.3/M input · $0.9/M output
Deployment
Cloud API

Our take

Codestral 2508 from Mistral offers a 256K-token context window and is well-suited to coding, long-context work. Pricing is $0.3/M input · $0.9/M output, billed per token through the provider API.

It is a balanced pick with no major caveats for most use cases. License: mistral-ai-non-production.

Compare to

DeepSeek V4 Flash
DeepSeek
1.05M context · $0.09/M input · $0.18/M output
Gemini 3.1 Flash Lite
Google
1.05M context · $0.25/M input · $1.5/M output
Llama 4 Scout
Meta
10M context · $0.1/M input · $0.3/M output

Not sure this is the one?

Describe your project and let LMFinder rank the field for your exact constraints.

Describe your project →