Recursive launches Flint, a laptop coding model built largely with AI
Current overview (not historical)
Recursive has released Flint, a coding model designed to run on laptops and largely built with AI assistance. Flint is a mixture-of-experts model that activates 2.5 billion parameters per task and is free to download on Hugging Face under the Apache 2.0 license for offline use and code privacy. In the SWE-Bench benchmark among models with fewer than 50 billion parameters, Flint ranked just behind Qwen2.5 27B in Recursive's six-item average score.
Written with AI from the reporting · updated 10/21/2026
Expand the detailed fact route and sources
Story route
A station-by-station route from the first public evidence onward; every station carries its own evidence and uncertainty.
- 1Unclassified fact stationOct 21, 00:00recorded Oct 22, 04:53single sourcerevised
Recursive launches Flint, a laptop coding model built largel
Read original source ↗ TechCrunch · AIReplay this moment in the explorerEvidence & revisions · 1 source records
- Subject：
- Recursive
- Action：
- launched Flint
- Object：
- coding model
- Occurred (fact side)：
- Oct 21, 00:00
- Published (valid side)：
- Oct 22, 04:27
- First seen (known side)：
- Oct 22, 04:53
Fact stationSingle source · one-sided claimRevisedCopy may have been revised later; time replay is bounded. Layer labels group reading only, not causation between layers.
Uncertainty marks flag where to check the sources; they are not proven causation.
Report timeline
Explore every side through dated source reports.
- TechCrunch · AIRelated fact: single sourceSource report revisedRecursive launches Flint, a laptop-run coding model largely built by AI
Report first recorded Oct 22, 04:53
Recursive, an AI startup, released Flint, a free coding model capable of running on a laptop. Built largely with its automated research system, the Eureka Machine, in under four months, Flint is a mixture-of-experts model using only 2.5 billion of its 40 billion parameters at a time. It is available on Hugging Face under the Apache 2.0 license, allowing offline use and code privacy. Flint demonstrated strong performance, outperforming other models under 50 billion parameters on SWE-Bench Verified, and achieved high token generation speeds on various hardware, though it requires more guidance than large paid assistants.
Read original source ↗
Event discussion
Reader opinions are hosted on GitHub Discussions; posting requires a GitHub account. Comments are not verified facts.