---
format: "aidr-story-markdown/v1"
id: "36d25b6fc71e100034eef6f095d28a76c7d945bd73c0e9365ec222750a8ba788"
canonical_url: "https://aidr.today/36d25b6f?lang=en"
title: "Cognition Launches SWE-2 Model at 70% Lower Cost for Devin Voice"
lang: "en"
requested_lang: "en"
available_langs: ["en","vi"]
translation_fallback: null
fallback_fields: []
published_at: "2026-09-10T23:46:50.000Z"
category: "Models"
topics: ["cognition","devin","llm","coding","agent"]
source_urls: ["https://huggingnews.com/ai/update-cognition-launches-swe-2-model-at-70percent-lower-cost-for-devin-eca4d736","https://x.com/cognition/status/2098142686486356185","https://x.com/cognition/status/2098142688654524430","https://x.com/cognition/status/2098069235733823965","https://x.com/cognition/status/2098069247842812103","https://x.com/cognition/status/2098069238044860487","https://x.com/cognition/status/2098069241039638917","https://x.com/cognition/status/2098069244151791618"]
summary: "Spoken commands are now integrated into the Devin AI software engineer via a voice interface that leverages GPT-Live and a new underlying model called SWE-2. This model performs on par with current frontier systems but at a cost up to 70% lower, utilizing reinforcement learning scaled to trillions of parameters. The update is currently available in the Devin Desktop and CLI applications. Post-trained on Kimi-K3, SWE-2 achieved a 50.0% score on the FrontierCode benchmark, beating GPT 5.6 Sol and Grok 4.6. The model includes effort levels where the medium setting takes 58% fewer turns and costs 81% less than the prior SWE-1.7 iteration. Cognition is offering the model free to Pro, Max, and Teams subscribers for one month."
---

# Cognition Launches SWE\-2 Model at 70% Lower Cost for Devin Voice

> [Open the canonical story](<https://aidr.today/36d25b6f?lang=en>)

**Published:** 2026-09-10T23:46:50.000Z
**Category:** Models
**Topics:** cognition, devin, llm, coding, agent

## Summary

Spoken commands are now integrated into the Devin AI software engineer via a voice interface that leverages GPT\-Live and a new underlying model called SWE\-2\. This model performs on par with current frontier systems but at a cost up to 70% lower, utilizing reinforcement learning scaled to trillions of parameters\. The update is currently available in the Devin Desktop and CLI applications\. Post\-trained on Kimi\-K3, SWE\-2 achieved a 50\.0% score on the FrontierCode benchmark, beating GPT 5\.6 Sol and Grok 4\.6\. The model includes effort levels where the medium setting takes 58% fewer turns and costs 81% less than the prior SWE\-1\.7 iteration\. Cognition is offering the model free to Pro, Max, and Teams subscribers for one month\.

## Sources

- [Story source](<https://huggingnews.com/ai/update-cognition-launches-swe-2-model-at-70percent-lower-cost-for-devin-eca4d736>)
- [Story source](<https://x.com/cognition/status/2098142686486356185>)
- [Supporting source](<https://x.com/cognition/status/2098142688654524430>)
- [Story source](<https://x.com/cognition/status/2098069235733823965>)
- [Story source](<https://x.com/cognition/status/2098069247842812103>)
- [Supporting source](<https://x.com/cognition/status/2098069238044860487>)
- [Supporting source](<https://x.com/cognition/status/2098069241039638917>)
- [Supporting source](<https://x.com/cognition/status/2098069244151791618>)