Digital Event Horizon
In a move that could significantly shift the landscape of AI coding, French startup Mistral has released its latest open-weights coding model, Devstral 2. This new development follows a year of significant advancements in the field of artificial intelligence and marks a shift towards more accessible and affordable solutions for developers. With its ability to maintain coherency across entire projects and detect failures, this model is seen as a serious contender in the market for AI-powered development tools.
Mistral has released Devstral 2, an open-weights coding model, with pricing starting at $0.40 per million input tokens. Devstral Small 2, a smaller-scale version of the model, costs $0.10 per million input tokens. Devstral 2 has achieved a score of 72.2 percent on SWE-bench Verified, a benchmark for testing AI systems. Mistral Vibe is a new development app that allows developers to interact with Devstral models directly in their terminal. Devstral 2 offers more control over projects and mitigates concerns about relying on AI-generated code.
In a move that is being closely watched by the AI research community, French startup Mistral has released its latest open-weights coding model, Devstral 2. This new development follows a year of significant advancements in the field of artificial intelligence and marks a shift towards more accessible and affordable solutions for developers.
According to reports from sources within the industry, Devstral 2 is currently being offered for free through Mistral's API, with pricing set to kick in once the free period expires. The cost structure is as follows: $0.40 per million input tokens and $2.00 per million output tokens. In contrast, Devstral Small 2, a smaller-scale version of the model with 24 billion parameters, costs $0.10 per million input tokens and $0.30 per million output tokens.
This pricing strategy is seen as a strategic move by Mistral to undercut its proprietary rivals in the market for AI coding models. According to reports, Devstral 2 has already achieved a score of 72.2 percent on SWE-bench Verified, a benchmark that attempts to test whether AI systems can solve real GitHub issues. This achievement puts the model squarely among the top-performing open-weights models.
The release of Devstral 2 also comes with the announcement of a new development app called Mistral Vibe. This command line interface (CLI) is designed to let developers interact with the Devstral models directly in their terminal and allows for tasks such as scanning file structures, detecting failures, retrying with corrections, tracking framework dependencies, and even handling bug fixing and modernizing legacy systems at repository scale.
The release of Devstral 2 has significant implications for the future of AI coding. With its ability to maintain coherency across entire projects and detect failures, this model is seen as a serious contender in the market for AI-powered development tools. While some have expressed concerns about the potential risks associated with relying on AI-generated code, Mistral's approach aims to mitigate these concerns by providing developers with more control over their projects.
The release of Devstral 2 and Mistral Vibe also highlights the evolving landscape of AI research and its applications in software development. As AI continues to advance at a rapid pace, it is likely that we will see even more innovative solutions emerge in the coming years. For now, the introduction of this new model marks an important milestone in the ongoing quest for more efficient and effective AI-powered coding tools.
In addition to Devstral 2, Mistral also released Devstral Small 2, a 24 billion parameter version that scores 68 percent on the same benchmark and can run locally on consumer hardware like a laptop with no Internet connection required. Both models support a 256,000 token context window, allowing them to process moderately large codebases.
While some have expressed concerns about the potential risks associated with relying on AI-generated code, Mistral's approach aims to mitigate these concerns by providing developers with more control over their projects. As the field of AI continues to advance, it is likely that we will see even more innovative solutions emerge in the coming years.
The release of Devstral 2 and its accompanying development app marks an important milestone in the ongoing quest for more efficient and effective AI-powered coding tools. With its ability to maintain coherency across entire projects and detect failures, this model is seen as a serious contender in the market for AI-powered development tools.
Related Information:
https://www.digitaleventhorizon.com/articles/A-New-Era-in-AI-Coding-Mistrals-Devstral-2-Model-Takes-on-Proprietary-Rivals-deh.shtml
https://arstechnica.com/ai/2025/12/mistral-bets-big-on-vibe-coding-with-new-autonomous-software-engineering-agent/
https://bardai.ai/2025/12/10/a-brand-new-open-ai-coding-model-is-closing-in-on-proprietary-options/
Published: Wed Dec 10 16:43:01 2025 by llama3.2 3B Q4_K_M
