token.app › Compare › Codestral 2508 vs Mercury 3
Codestral 2508 vs Mercury 3
Mercury 3 is 2.8× cheaper per token on a blended 3:1 basis. Mercury 3 accepts the larger context window (1M).
Head to head
Blended price uses a 3:1 input:output mix
| Codestral 2508 | Mercury 3 | |
|---|---|---|
| Provider | Mistral AI | Inception |
| Input $/1M | $0.300 | $0.100 |
| Output $/1M | $0.900 | $0.350 |
| Blended $/1M | $0.450 | $0.163 |
| Context window | 256K | 1M |
| Max output | 205K | 131K |
| Released | Aug 2025 | Oct 2026 |
| Licence | Proprietary | Proprietary |
Choose Codestral 2508 if…
- No clear advantage on the data we hold.
Choose Mercury 3 if…
- Costs less per token — $0.163 vs $0.450 blended
- Larger context window (1M)
- Newer model (released Oct 2026)
FAQ
Which is better, Codestral 2508 or Mercury 3?
We do not hold independently-run benchmark scores covering both Codestral 2508 and Mercury 3, so we make no quality claim. Their pricing and specifications are compared on this page.
Is Codestral 2508 cheaper than Mercury 3?
Mercury 3 is cheaper: $0.163 versus $0.450 per million tokens blended at 3:1 input:output. Input alone: $0.300 vs $0.100. Output alone: $0.900 vs $0.350.
What context windows do Codestral 2508 and Mercury 3 have?
Codestral 2508: 256,000 tokens. Mercury 3: 1,000,000 tokens.
Who makes Codestral 2508 and Mercury 3?
Codestral 2508 is made by Mistral AI. Mercury 3 is made by Inception.