Model comparison
Mistral Large vs DeepSeek R1
Mistral
Mistral Large
European sovereignty and disciplined function calling.
DeepSeek
DeepSeek R1
The open-weights reasoning model that reset the cost curve.
Specs
| Metric | Mistral Large | DeepSeek R1 |
|---|---|---|
| Context window | 128K tokens↑ | 64K tokens |
| Input $/1M tokens | $2.00 | $0.700↑ |
| Output $/1M tokens | $6.00 | $2.50↑ |
| Modalities | Text | Text |
| Open weights | No | Yes |
| Released | - | Jan 2025 |
Capability differences
| Capability | Mistral Large | DeepSeek R1 |
|---|---|---|
| Tool use | Yes | No |
| Extended thinking | No | Yes |
| Open weights | No | Yes |
Key differences
- -Mistral Large offers 128K context; DeepSeek R1 offers 64K.
- -Input pricing: Mistral Large at $2.00/1M vs DeepSeek R1 at $0.700/1M.
- -DeepSeek R1 is open weights; Mistral Large is proprietary.