GLM-4-9B: open-source model with superior performance to Llama-3-8B

Looks like terrific technology. However, the translation says that it's an "irrevocable revocable" non-commercial license with a form to apply for commercial use.

I’m excited to hear work is being done on models that support function calling natively.<p>Does anybody know if performance could be greatly increased if only a single language was supported ?<p>I suspect there’s a high demand for models that are maybe smaller and can run faster if the tradeoff is support for only English.<p>Is this available in ollama ?

&gt; GLM-4V-9B possesses dialogue capabilities in both Chinese and English at a high resolution of 1120*1120. In various multimodal evaluations, including comprehensive abilities in Chinese and English, perception &amp; reasoning, text recognition, and chart understanding, GLM-4V-9B demonstrates superior performance compared to GPT-4-turbo-2024-04-09, Gemini 1.0 Pro, Qwen-VL-Max, and Claude 3 Opus.<p>But according to their own evaluation further down, gpt-4o-2024-05-13 outperforms GLM-4V-9B on every task except OCRBench.

Isnt 3-70b so good, reddit llamaers are saying people should buy hardware to run it?<p>Llama-3-8b was garbage for me but damn 70b is good enough

If those numbers are true then it&#x27;s very impressive. Hoping for llama.cpp support.

1M context, but does it really? I&#x27;ve been hit with 32K models that crap out after 10K before...

model available, not open source.

Ehhh man this is frustrating, 7B was a real sweet spot for hobbyist. 8B...doable. I&#x27;ve been joking to myself&#x2F;simultaneously worried that Llama 3 8B and Phi-3 &quot;3B&quot; (3.8B) would start a &quot;ehhh, +1, might as well be a rounding error&quot; thing. It&#x27;s a big deal! I measure a 33% decrease just going from 3B to 3.8B when inferencing on CPU.