SWE-1.7
The most capable model Cognition has trained so far. It reaches frontier-level intelligence at a much lower cost, advancing the cost-performance Pareto curve.
Specs
Model
- Family
- SWE
- Variant
- 1.7
- Weights
- Proprietary
Modalities
- Input
- Text
- Output
- Text
Capabilities
- Tool calling
- Yes
Dates
- Released
- 8 Jul 2026