Mistral's 22B code-specialized dense LLM under Mistral Non-Production License; supports 80+ programming languages, fill-in-the-middle and code completion.
@hn_2e7430
about 1 month ago
I found Mistral's agentic coding models to be fast and good, but not always accurate. I think the reason for that is that most IDEs aren't properly setup to use their models. But I haven't tried this new Codestral version.
@hn_d76ad2
about 1 month ago
Also looks like an older version of Codestral works well with TabbyML: https://tabby.tabbyml.com/blog/2024/07/09/tabby-codestral/ Thank you for sharing, this is almost exactly what I've been looking for, for ages!
@hn_e2441d
about 1 month ago
Also worth checking out was codestral... I think that had a 256k context and used Mamba even if it is slightly older model now... it had worked great for a Text2SQL use case we worked on.
@hn_239000
2 months ago
a bit of surprise since codestral is among best open models so far.
@hn_3d27c7
2 months ago
But at the same time they don't open weights of Codestral