Cognition Launches Affordable Plan for AI Coding Tool Devin
Cognition, the company behind the popular AI-powered coding assistant Devin, has unveiled a new budget-friendly pricing tier to attract more users, as reported by TechCrunch. Devin made headlines last year for its ability to autonomously handle certain software development tasks, quickly going viral across social media platforms.
However, while it gained attention, developers soon discovered that Devin struggled with more complex coding challenges. Despite its limitations, the tool earned praise from tech leaders such as Perplexity CEO Aravind Srinivas, significantly boosting Cognition’s visibility in the competitive AI market.
From Premium-Only to Accessible Pricing
TechCrunch reports that initially, Devin was offered exclusively to teams at a steep $500 monthly subscription. But following Cognition’s recent multimillion-dollar funding round, the company has introduced a more accessible $20 entry-level option. After the initial payment, users shift to a pay-as-you-go model.
However, users should be mindful of costs under the new plan. The $20 fee provides about nine ACUs, Cognition’s internal currency for computing credits. Each ACU, priced at $2.25 in this lower-tier plan (up from $2 in the premium subscription), equates to around 15 minutes of Devin’s active work. This means the initial package covers roughly 2.25 hours of usage, a potentially short span for handling large code projects.
Also read: AI Video Startup Runway Raises $308 Million to Grow New Media Businesses
Enhanced Capabilities in Devin 2.0
Cognition asserts that Devin’s latest version, Devin 2.0, is substantially improved from its initial release. Inspired by tools like GitHub Copilot, Devin now assists with project planning, provides code explanations complete with citations, and can generate documentation to build "wikis" for codebases.
Silas Alberti, part of the Devin development team, told TechCrunch that the tool now delivers "twice as much work" as it did previously.
Performance Still Under Scrutiny
While the company touts these enhancements, experts advise cautious optimism. Studies consistently show that even the most advanced code-generating AI tools can inadvertently introduce bugs and security risks due to gaps in understanding complex programming logic. In fact, a recent assessment of Devin revealed that it completed only three out of 20 assigned tasks successfully, highlighting the challenges that still remain.
