← Catalog
SWE-2 Flash
AvailableCognitionProprietary
Cognition's smaller, faster coding model, released Nov 4 2026 in Devin Desktop and CLI at one-fifth the price and 3x the speed of SWE-2. Built on Thinking Machines' Inkling-Small base with a new mid-training stage before RL; Cognition reports Terminal-Bench 4.0 rising from 0.9% (base) to 20.0% and FrontierCode 2.0 from 3.2% to 24.0%, plus ExploitBench 34.8%. Weights are not public; Cognition plans a program to share them with vetted partners and researchers. Benchmarks vendor-reported.
Specifications
- License
- Proprietary
- Weights
- Not released
- Architecture
- unknown
- Parameters
- Undisclosed
- Context window
- — tokens
- Max output
- —
- Knowledge cutoff
- —
- Price (in / out, $/M)
- —
- Modalities
- TextCode
Benchmarks
No benchmark scores recorded yet. Spotted some? Submit a correction.
Vendor-reported figures are claims until independently verified. See methodology.