Cognition's SWE-2 logo

Cognition's SWE-2

Cognition's coding model, 64% cheaper than Fable 5.1

Listing details

HiCyou directory record

Listed
Listed website
cognition.com
Directory record created
Sep 14, 2026
Last record update
Sep 14, 2026

Listing information

  • Overview
  • 5 key features
  • 3 use cases

This page describes HiCyou's directory record. It is not a security, ownership, or product-quality certification.

Review the external site before sharing sensitive information or making a purchase.

Cognition's SWE-2 preview

What is Cognition's SWE-2

SWE-2 is Cognition's coding model, post-trained from Kimi K3 with reinforcement learning that optimizes for cost and capability simultaneously. According to Cognition, it scores 50.0% on FrontierCode 1.1 Main, within a point of Fable 5.1 at 64% lower cost, and near GPT-6 Astra at a quarter of the cost. Compared to SWE-1.7, it takes 58% fewer turns, costs 81% less, and scores higher. It is available now in Devin Desktop and CLI.

Key Features

Post-trained from Kimi K3 using RL optimizing for cost and capability
Scores 50.0% on FrontierCode 1.1 Main
Claimed 64% cheaper than Fable 5.1 with comparable performance
58% fewer turns and 81% lower cost versus SWE-1.7 with higher scores
Available in Devin Desktop and Devin CLI

Use Cases

  • Agentic coding tasks within Devin Desktop
  • Software engineering automation via Devin CLI
  • Cost-efficient code generation and iteration for development teams

Why do startups need this tool?

SWE-2 targets the cost-performance frontier for coding models, offering near frontier-level agentic coding performance at a fraction of the price. For startups running AI-assisted development workflows, it promises fewer turns and lower costs compared to its predecessor, and it integrates directly into Devin Desktop and CLI.

FAQs

Cognition's SWE-2 Alternatives

Fable 5.1
GPT-6 Astra
SWE-1.7