SWE-2
5.0 out of 5 starsAverage Ratings 0 Ratings
Average Ratings 1 Rating
Description
MAI Code 2 Flash is an AI-powered coding model designed to deliver fast, affordable assistance for routine software development and programming workflows. Developed by Microsoft AI, the model belongs to the MAI Code 2 family and provides a smaller, more efficient alternative to the larger MAI Code 2 model. Its architecture combines a decoder-only transformer with sparse Mixture-of-Experts layers and an integrated vision encoder, using approximately 12 billion active parameters. The model supports a context window of 1.05 million tokens, allowing it to process extensive source code, technical documentation, and other development materials. MAI Code 2 Flash specializes in explaining existing code, implementing targeted modifications, fixing common programming errors, and supporting rapid development iterations. Its agentic coding capabilities also support repository analysis, multi-file editing, debugging, and interaction with development tools. Configurable reasoning effort settings allow developers to balance response quality, processing speed, and inference costs according to task requirements. According to Microsoft's published benchmarks, MAI Code 2 Flash achieves performance comparable to GPT-6 Luna and Claude Haiku 5.5 at substantially lower costs under specified evaluation conditions. With availability through GitHub Copilot and Microsoft Foundry, the model helps software engineers and enterprises automate frequent coding activities and improve development efficiency.
Description
SWE-2 is a software engineering model from Cognition built for agentic coding tasks that require strong performance at lower computational and monetary cost. It is post-trained from the Kimi K3 base model and extends Cognition’s earlier SWE-1.7 training approach with a new reinforcement learning method for jointly optimizing multiple reasoning-effort settings. Medium, high, and maximum effort modes provide different tradeoffs between speed, cost, exploration, and verification depending on task complexity. The model is trained to inspect only the parts of a codebase that are likely to matter, helping it reach implementation faster and reduce unnecessary exploration. SWE-2 can generate and modify code, run tests, analyze repositories, work through terminal tasks, and verify whether implementations satisfy user requirements. Cognition also reports improvements in end-to-end test creation, regression detection, instruction following, and re-deriving conclusions when challenged. Its training process incorporates cost-aware rewards, length-weighted reward baselines, expanded reinforcement learning environments, and hardened verifiers intended to improve both efficiency and reliability. SWE-2 is positioned as a cost-efficient alternative to larger frontier coding models while remaining competitive on software engineering benchmarks such as FrontierCode, DeepSWE, and Terminal-Bench. The model is available in Devin Desktop and Devin CLI and is being introduced to additional Cognition products including Devin Web and Fusion.
API Access
Has API
Yes
API Access
Has API
No
Integrations
.NET
Yes
C#
No
CSS
No
Cerebras
No
Dart
No
Devin Desktop
No
Go
No
JSON
No
JavaScript
No
Kotlin
No
Integrations
.NET
Yes
C#
Yes
CSS
Yes
Cerebras
Yes
Dart
Yes
Devin Desktop
Yes
Go
Yes
JSON
Yes
JavaScript
Yes
Kotlin
Yes
Pricing Details
$0.05 per 1M tokens (input)
$0.05 per 1M input tokens, $0.005 per 1M cached input tokens, and $0.25 per 1M output tokens
Free Trial
No
Free Version
No
Pricing Details
$20/month
Free Trial
No
Free Version
Yes
Deployment
Web-Based
Yes
On-Premises
No
iPhone App
No
iPad App
No
Android App
No
Windows
No
Mac
No
Linux
No
Chromebook
No
Deployment
Web-Based
Yes
On-Premises
No
iPhone App
No
iPad App
No
Android App
No
Windows
No
Mac
No
Linux
No
Chromebook
No
Customer Support
Business Hours
No
Live Rep (24/7)
No
Online Support
Yes
Customer Support
Business Hours
No
Live Rep (24/7)
No
Online Support
Yes
Types of Training
Training Docs
Yes
Webinars
No
Live Training (Online)
No
In Person
No
Types of Training
Training Docs
Yes
Webinars
No
Live Training (Online)
No
In Person
No
Vendor Details
Company Name
Microsoft AI
Founded
2024
Country
United States
Website
microsoft.ai/
Vendor Details
Company Name
Cognition
Founded
2023
Country
United States
Website
cognition.com