Mistral: Devstral Small (free)
Devstral-Small-2505 is a 24B parameter agentic LLM fine-tuned from Mistral-Small-3.1, jointly developed by Mistral AI and All Hands AI for advanced software engineering tasks. It is optimized for codebase exploration, multi-file editing, and integration into coding agents, achieving state-of-the-art results on SWE-Bench Verified (46.8%).
Devstral supports a 128k context window and uses a custom Tekken tokenizer. It is text-only, with the vision encoder removed, and is suitable for local deployment on high-end consumer hardware (e.g., RTX 4090, 32GB RAM Macs). Devstral is best used in agentic workflows via the OpenHands scaffold and is compatible with inference frameworks like vLLM, Transformers, and Ollama. It is released under the Apache 2.0 license.
131,072 Token Context
Process and analyze large documents and conversations.
Advanced Coding
Improved capabilities in front-end development and full-stack updates.
Agentic Workflows
Autonomously navigate multi-step processes with improved reliability.
Available On
| Provider | Model ID | Context | Max Output | Input Cost | Output Cost | Throughput | Latency |
|---|---|---|---|---|---|---|---|
| DeepInfra | deepInfra | 128K | - | $0.06/M | $0.12/M | 53.5 t/s | 502 ms |
| NextBit | nextBit | 131K | - | $0.07/M | $0.10/M | 42.5 t/s | 1186 ms |
| Parasail | parasail | 131K | 131K | $0.09/M | $0.30/M | 57.9 t/s | 877 ms |
| Mistral | mistral | 131K | - | $0.10/M | $0.30/M | 113.5 t/s | 409 ms |