MMistral · Released Dec 9, 2025
Devstral 2 2512
Devstral 2 is a state-of-the-art open-source model by Mistral AI specializing in agentic coding. It is a 123B-parameter dense transformer model supporting a 256K context window. Devstral 2 supports exploring...
- Released
- Dec 9 2025
- Context window
- 262K 210K out
- Input / 1M tokens
- $0.40 from Free
- Output / 1M tokens
- $2.00 from Free
Where to run it
14 providers compared
Free to try on InferX — free endpoints are often limited-time and rate-limited. Cheapest: NanoGPT at $0.40 / $1.40 per 1M input / output tokens — 46% less than the priciest option.
| Provider | Input | Output | Cache read | Context | Max output | Uptime | vs. best |
|---|---|---|---|---|---|---|---|
| InferXfree · may be limited-time | Free | Free | — | 128K | 128K | — | Free |
| NanoGPTno tool calling | $0.40 | $1.40 | $0.20 | 262K | 66K | — | Cheapest |
| Mistraleu | $0.40 | $2.00 | $0.04 | 262K | 210K | 100%24h | +23% |
| LLM Gatewayno tool calling | $0.40 | $2.00 | — | 262K | 262K | — | +23% |
| Opper | $0.40 | $2.00 | — | 256K | 8K | — | +23% |
| Merge Gateway | $0.40 | $2.00 | $0.04 | 256K | 256K | — | +23% |
| Kilo Gateway | $0.40 | $2.00 | $0.04 | 262K | 210K | — | +23% |
| DevPass (LLM Gateway) | $0.40 | $2.00 | — | 262K | 262K | — | +23% |
| Amazon Bedrock | $0.40 | $2.00 | — | 262K | 8K | — | +23% |
| Pioneer | $0.40 | $2.00 | $0.40 | 256K | 131K | — | +23% |
| Eden AI | $0.40 | $2.00 | $0.04 | 262K | 262K | — | +23% |
| Cortecs | $0.48 | $2.39 | $0.045 | 256K | 256K | — | +47% |
| GreenPT | $0.57 | $2.74 | — | 200K | 16K | — | +71% |
| AI/ML API | $0.60 | $3.03 | $0.061 | 262K | 262K | — | +86% |
Benchmarks
Benchmark scores
No benchmark scores yet — Epoch AI hasn't evaluated Devstral 2 2512. Scores appear here automatically once it does.
Specs
Details
- Model ID
mistralai/devstral-2512- Developer
- Mistral
- Release date
- Dec 9, 2025 ·
- Context window
- 262,144 tokens
- Max output
- 209,715 tokens
- API price
- $0.40 / $2.00 per 1M input / output tokens
- Providers
- 14
- Input types
- text, file
- Reasoning
- No
- Tool calling
- Yes
- Open weights
- Yes
- Free tier
- No
Release date and prices from OpenRouter's catalog; providers from OpenRouter and models.dev. Buying directly from a provider can cost a little differently. Speed, latency and uptime are measured on OpenRouter only; uptime is the share of OpenRouter requests each provider served over the last day. Benchmark scores are the best result per benchmark from Epoch AI, licensed CC BY 4.0. Rank counts every model tracked here with a score.