token.app › Compare › Codestral 2508 vs Kimi K3
Codestral 2508 vs Kimi K3
Codestral 2508 is 13× cheaper per token on a blended 3:1 basis. Kimi K3 accepts the larger context window (1.0M).
Head to head
Blended price uses a 3:1 input:output mix
| Codestral 2508 | Kimi K3 | |
|---|---|---|
| Provider | Mistral AI | Moonshot AI |
| Input $/1M | $0.300 | $3.00 |
| Output $/1M | $0.900 | $15.00 |
| Blended $/1M | $0.450 | $6.00 |
| Context window | 256K | 1.0M |
| Max output | — | — |
| Released | Aug 2025 | Jul 2026 |
| Licence | Proprietary | Open weights |
Choose Codestral 2508 if…
- Costs less per token — $0.450 vs $6.00 blended
Choose Kimi K3 if…
- Larger context window (1.0M)
- Open weights — self-hostable
- Newer model (released Jul 2026)
FAQ
Which is better, Codestral 2508 or Kimi K3?
We do not hold independently-run benchmark scores covering both Codestral 2508 and Kimi K3, so we make no quality claim. Their pricing and specifications are compared on this page.
Is Codestral 2508 cheaper than Kimi K3?
Codestral 2508 is cheaper: $0.450 versus $6.00 per million tokens blended at 3:1 input:output. Input alone: $0.300 vs $3.00. Output alone: $0.900 vs $15.00.
What context windows do Codestral 2508 and Kimi K3 have?
Codestral 2508: 256,000 tokens. Kimi K3: 1,048,576 tokens.
Who makes Codestral 2508 and Kimi K3?
Codestral 2508 is made by Mistral AI. Kimi K3 is made by Moonshot AI.