AI startup Recursive is starting to live up to its name. The company, which is trying to build AI that can improve itself, released its first model on Wednesday: Flint, a free coding model that runs on a laptop — and was, it says, largely built by AI.
Recursive (Credits: TechCrunch / Getty Images)
An AI That Helped Build Itself
Recursive’s automated research system, which it calls the Eureka Machine, contributed to nearly every step of Flint’s development, the company says — from pre-training to post-training and inference. That allowed a small team to build the model from scratch in under four months.
Flint is an early public test of recursive self-improvement: the idea that AI can build better AI, which in turn builds even better AI. It’s one of the most debated ideas in the industry — some researchers fear self-improving AI could slip out of human control. For now, Recursive’s version is more limited: an AI research assistant overseen by people that produced a model small enough for a laptop.
The Eureka Machine comes up with ideas, runs experiments, and uses the results to decide what to try next. The company credits it with a sparser model shape, low-bit training that shrank Flint by 71% with almost no loss in quality, and custom kernels — specialized programs that speed up the underlying computations. Recursive isn’t alone in this trend: OpenAI said last month that by mid-August, its AI agents were logging 3.1 workdays for every one worked by a human researcher.
A Tiny, Blazing-Fast Coding Model
Because Flint runs on the user’s own computer rather than a company’s cloud servers, developers can use it offline, keep their code private, and avoid the usage fees AI cloud services charge. The model is free to download on Hugging Face under the Apache 2.0 open-source license.
Flint is a mixture-of-experts model: it uses only 2.5 billion of its 40 billion parameters at a time, making it light enough for an ordinary laptop. In Recursive’s tests, it generated about 131 tokens per second on a MacBook Pro with Apple’s M4 Pro chip — roughly twice the 58 to 69 tokens per second of similar models like Google’s Gemma 4 and Nvidia’s Nemotron 3.5 Lightning on the same machine. On Nvidia’s RTX Spark, it topped 150 tokens per second.
On SWE-Bench Verified, a benchmark measuring how well AI fixes real-world software bugs, Flint outperformed every other model under 50 billion parameters. Early hands-on reviews were positive too: Wired’s Will Knight used Flint on a Mac with 48GB of memory to build several small programs, including a to-do list app and a web research tool, finding that Flint “nailed his coding requests every time” while other small models ran into errors or lagged.
The Startup Behind It
Recursive, founded last year and led by CEO Richard Socher, Salesforce’s former chief scientist, emerged from stealth in May with $650 million in funding at a $4.65 billion valuation, from backers including GV, Nvidia, and AMD Ventures.
The company acknowledged Flint would need more guidance than the big paid coding assistants, and suggested breaking large tasks into smaller steps. But as an early demo of AI building better AI, it’s one to watch.
Also Read: While Amazon Blocks AI Agents, Walmart Rolls Out Agent Gateway to Let AI Shop for You
Source: TechCrunch — read the full story there.
