Devstral 2512 (listed on OpenRouter as "Devstral 2") is Mistral AI's open-source agentic coding model. It is a 123-billion-parameter dense transformer. It is available via OpenRouter. The price listed here is the OpenRouter listed price, checked Nov 18, 2026: $0.40 per million input tokens and $2.00 per million output tokens. Data Source: OpenRouter listing. Data Confidence: Partial.
Devstral 2512 supports a 262,144-token context window and a maximum output of 209,715 tokens. It supports function calling and streaming. It does not support vision input. It is open-weight and is not classified as a reasoning model on this page.
Within the Devstral family, Devstral 2512 is the 123B dense transformer variant. Devstral Medium and Devstral Small 2505 are the other members of the family. At $0.40/$2.00 per million tokens, Devstral 2512 is priced identically to Devstral Medium via OpenRouter.
Data source: OpenRouter listing, checked Nov 18, 2026. Data Confidence: Partial — throughput, latency, and benchmark scores are not available for this model at this time. The OpenRouter listed price may differ from Mistral AI's direct API pricing.
Devstral 2512 is suited to agentic coding workflows that benefit from a large open-weight model with a 262K context window. Its open-weight availability makes it suitable for self-hosted deployments where data privacy or customization is a requirement.
Input price vs 191 advanced models tracked. Avg: $1.37/1M.
| Model ID | devstral-2512 |
| Provider | Mistral |
| Model Family | Devstral |
| Tier | Advanced |
| Context Window | 262,144 tokens (262K) |
| Max Output Tokens | 209,715 tokens |
| Input Price | $0.4 / 1M tokens |
| Output Price | $2 / 1M tokens |
| Cached Input Price | Not available |
| Throughput | — |
| Latency (P50) | — |
| Intelligence Score | — |
| LiveBench Score | — |
| Cost per Task (LiveBench) | — |
| Data Confidence | Partial |
| HLE Score | — |
| GPQA Diamond | — |
| SWE-bench Score | — |
| OSWorld Score | — |
| BrowseComp Score | — |
| Terminal-Bench Score | — |
| AutoBench Score | — |
| Parameters | 123B |
| Vision Support | No |
| Function Calling | Yes |
| Streaming | Yes |
| Reasoning Model | No |
| Open Weights | Yes |
| Data Source | Static (verified) |
| Last Updated | 2026-11-18 |
Self-hosting devstral-2512? These GPUs are well-matched for this model's compute and memory requirements.