Cognition Launches SWE-1.7, Frontier-Level AI Model at Lower Cost
Cognition has released SWE-1.7, a highly capable AI model that achieves frontier-level intelligence at a lower cost than previous models. The model was trained using a combination of better infrastructure, more stable training, higher-quality data, and new techniques. SWE-1.7 is available for use via Cerebras at 1000 transactions per second.
Key points
- Cognition, a tech company, has launched SWE-1.7, a highly capable AI model that achieves frontier-level intelligence at a lower cost.
- The model was trained using a combination of better infrastructure, more stable training, higher-quality data, and new techniques.
- SWE-1.7 is particularly optimized for longer-horizon asynchronous tasks, an important component of high-quality software engineering.
- The model has been made available for use via Cerebras at 1000 transactions per second.
- Cognition has been formulating and refining principles for good agentic software engineering, both in evaluation and in training, with SWE-1.7.
Cognition has made a significant breakthrough in AI research with the launch of SWE-1.7, a highly capable model that achieves frontier-level intelligence at a lower cost than previous models. The model was trained using a combination of better infrastructure, more stable training, higher-quality data, and new techniques.
SWE-1.7 is particularly optimized for longer-horizon asynchronous tasks, an important component of high-quality software engineering. This makes it a valuable tool for software development and engineering.
The model has been made available for use via Cerebras at 1000 transactions per second. Cognition has been formulating and refining principles for good agentic software engineering, both in evaluation and in training, with SWE-1.7.
This breakthrough has significant implications for the field of AI research and development. It demonstrates the potential for AI models to achieve high levels of intelligence at lower costs, making them more accessible and practical for use in a wide range of applications.
Sources
The WireByte editorial team synthesises technology news from multiple primary sources, verifies the facts, and links every source. Articles are produced with AI assistance and reviewed under our editorial policy.