Bluesky Thread

Starting with K2, several large “agentic coding” models weren’t trained as re...

View original thread
Starting with K2, several large “agentic coding” models weren’t trained as reasoners:

- K2
- GLM-4.5 (current SOTA open weights)
- Opus 4.1 benchmarked in non-thinking mode

seems like convergence on “reasoning isn’t necessary for coding”, or maybe undesirable since it uses up context
29
meanwhile, GPT-5 is going the opposite direction
11
29 likes 0 reposts

More like this

×