UNIT.City — місце, де люди працюють... КРАЩЕ! Обирай свій простір просто зараз 👉
Наталя ХандусенкоAI Eng
16 May 2025, 13:11
2025-05-16
Windsurf, a vibe-coding startup acquired by OpenAI for $3 billion, has released its first family of AI models for software engineering.
Windsurf, a startup that develops AI tools for software engineers, has announced the launch of its first family of AI models, SWE-1, short for AI software engineering models. These include SWE-1, SWE-1-lite, and SWE-1-mini, which are optimized for the entire software development process, not just coding, the startup claims.
Windsurf, a startup that develops AI tools for software engineers, has announced the launch of its first family of AI models, SWE-1, short for AI software engineering models. These include SWE-1, SWE-1-lite, and SWE-1-mini, which are optimized for the entire software development process, not just coding, the startup claims.
Windsurf is an AI startup that is being acquired by OpenAI. The deal is worth $3 billion, the largest in the history of the ChatGPT creator. The startup is known for its tools that allow software engineers to write and edit code through conversations with a chatbot.
Traditionally, vibe coding startups have relied on AI models from OpenAI, Anthropic, and Google. Windsurf seems to want to change that: with SWE-1, it can expand its activities not only to developing applications, but also to building the models that power them.
According to the startup, SWE-1 — the largest and most powerful AI model of the three — can compete with Claude 3.5 Sonnet, GPT-4.1, and Gemini 2.5 Pro in internal programming tests. However, it falls short of advanced AI models like Claude 3.7 Sonnet in software development tasks, TechCrunch reports .
“Today’s cutting-edge models are optimized for coding, and they’ve made significant strides in the last couple of years. But that’s not enough for us… Coding is not software engineering,” says Windsurf’s head of research, Nicholas Moy.
Windsurf notes in its blog that while other models are good at writing code, they struggle to work across multiple surfaces — as programmers often do — such as terminals, IDEs, and the internet. The startup says SWE-1 was trained using a new data model and “a training recipe that encapsulates unfinished states, long-running tasks, and multiple surfaces.”
The SWE-1-lite and SWE-1-mini models will be available to both paid and free users. Meanwhile, the SWE-1 will be available only to paid users. The startup hasn’t yet announced the exact price, but claims that it’s cheaper to maintain than the Claude 3.5 Sonnet.
“This leads to severe degradation and laziness.” Ukrainian programmers shared the secrets of effective vibe coding and the “side effects” of using AI when creating code
Як нейромережі бачать вільну та незалежну Україну? Тест dev.ua
Нейронні мережі для генерації зображень бачать світ по-своєму, їхню логіку зрозуміти часом зовсім неможливо. Але таки хочеться. На честь Дня Незалежності України редакція dev.ua вирішила провести невеликий експеримент.
Ми задали чотирьом різним нейронним мережам п’ять однакових запитів: «прапор України», «День Незалежності України», «український Крим», «перемога України» та «українці». Отриманими результатами ми ділимося з вами нижче.
У TikTok тепер можна генерувати фон за допомогою нейромережі. Ми протестували її та ділимося результатами
У TikTok з’явилася нова функція «Розумний фон». З її допомогою як фон для тіктоків можна підставляти згенеровані нейромережею зображення. Редакція dev.ua протестувала цю технологію і ділиться своїми враженнями.
https://upstaff.com/hire/full-stack-web/