Local coding models

RichardSocher introduces Flint with open weights for local coding

RichardSocher introduced Flint as @Recursive_SI's first model, a local coding model built with its automated research system, the Eureka Machine. The post said that open weights were available on Hugging Face and claimed speeds of 125+ tokens per second on a MacBook Pro and 150+ on RTX Spark.

tydsh reported scores of 54.9% on Terminal-Bench 2.1, 89.7% on SWE-bench Pro v2 and 74.2% on SWE-bench Verified, saying that the model was trained from scratch. UnslothAI said that its Dynamic GGUFs allow local use with 24GB RAM and described an Apache-2.0 model with 39.8B parameters, of which 2.5B are active.