On a lovely Thursday, Windsurf, a hotshot startup creating beloved AI tools for software engineers, blew us away with an announcement of the launch of their virgin family of AI software engineering models, fondly nicknamed SWE-1. The Windsurf family has proudly declared they've trained these new kids on the block to be ace performers through the "entire software engineering process," and not just be one-trick ponies at coding.

The unveiling of Windsurf's home-grown AI geniuses might have raised a few eyebrows, considering the rumor mill has been rife with talks of OpenAI closing the lock on a whopping $3 billion deal to gobble up Windsurf.The introduction of this new model, however, suggests that Windsurf intends to do more than just create applications since they understand how crucial it is to create the power source that drives them.

The story straight from the horse's mouth states that SWE-1, the mightiest member of the family, is giving a run for their money to formidable competitors like Claude 3.5 Sonnet, GPT-4.1, and Gemini 2.5 Pro in the internal programming arena. However, the mighty SWE-1 didn't quite reach the peak, with AI whizzes like Claude 3.7 Sonnet overtaking it on software engineering missions.

SWE-1-lite and SWE-1-mini, Windsurf's lesser champions, will be available to devotees on its powerful platform for free or with a subscription. Nevertheless, SWE-1 will only be used by subscribers. While Windsurf is yet to announce the price tag of its SWE-1 models, they make a bold statement, claiming it's light on the pocket compared to their adversary - Claude 3.5 Sonnet.

Windsurf shines in the spotlight for creating tools that empower software engineers to converse with an AI chatbot to draft and modify code. This savvy move, termed "vibe coding," is also championed by other startups like Cursor and Lovable. These businesses, like Windsurf, have historically relied on AI models that were sourced from Google, Anthropic, and OpenAI's tech experts.

A video that unveiled the SWE models was peppered with statements by Windsurf's think tank, Head of Research, Nicholas Moy, emphasising Windsurf's novel efforts to set themselves apart from the pack. "The current pioneers have tailored their models for coding, making leaps and bounds in recent years," declares Moy. "But, that's not our endgame...Coding isn't equivalent to software engineering."

Windsurf points out in a blog post that while competitors are top-notch code writers, they stumble when multitasking across various platforms, something software engineers frequently do, spanning terminals, IDEs, and the internet. SWE-1 was painstakingly trained using a new data model and a recipe that bottles various surfaces, long-winding activities, and partial states, according to Windsurf.