Mistral Large 3 2512 (batch)
Mistral · Multimodal · Released Dec 2025
Mistral Large 3 2512 is Mistral's most capable model, built on a sparse mixture-of-experts architecture with 41B active parameters and a 262K token context window.
- Strengths
- The model handles complex reasoning, long-context tasks, and nuanced language understanding across a wide range of domains.
- Best for
- Production workloads requiring frontier-level language capabilities, including batch processing, complex reasoning, and long-context document analysis.
- Limitations
- The sparse MoE design may introduce latency variability; for simple tasks or latency-critical applications, smaller models like Mistral Small 4 offer better trade-offs.
Input / 1M
$0.25
Output / 1M
$0.75
Cached input / 1M
$0.025
Context window
262K
Price history
Input (solid)Output (dashed)
Price change
- 30d
- in decreased 50.0% out decreased 50.0%
- 90d
- in decreased 50.0% out decreased 50.0%
- 1y
- in decreased 50.0% out decreased 50.0%
- Since launch
- in decreased 50.0% out decreased 50.0%
Snapshots
| Effective | Input | Output | Cached in | Note | Source |
|---|---|---|---|---|---|
| 10 Sep 2026 | $0.25 | $0.75 | $0.025 | Imported from OpenRouter | openrouter.ai |
| 28 Aug 2026 | $0.5 | $1.50 | $0.05 | Imported from OpenRouter | openrouter.ai |