model
Cognition launches new SWE-2 model, Rivaling Fable 5.1 and GPT-Astra
Overpromises on performance without concrete evidence.
3.5综合
实用性3
上手速度5
完成度4
不可替代3
生命力3
5 维独立打分，综合分为加权平均。分数只对同一套标准内部可比。
适合谁
Data scientists and AI researchers looking for a new model to experiment with.
不适合谁
Teams needing a reliable, battle-tested model for production.
项目自述
cognition.com
替代品
TensorFlowPyTorchHugging Face Transformers