12 ms·
With all due respect, you need to actually test out Qwen3 2507 or GLM 4.5 before making these sorts of claims. Both of them are comparable to OpenAI's largest m
by bigyabai 1y ago
With all due respect, you need to actually test out Qwen3 2507 or GLM 4.5 before making these sorts of claims. Both of them are comparable to OpenAI's largest models and even bench favorably to Deepseek and Opus: https://cdn-uploads.huggingface.co/production/uploads/62430a8522549d0917bfeb5a/0d7zztq4GB7G2ZYowO-dQ.jpeg https://cdn-uploads.huggingface.co/production/uploads/62430a...
It's cool to see OpenAI throw their hat in the ring, but you're smoking straight hopium if you think there's "no reason to run other open source models now" in earnest. If OpenAI never released these models, the state-of-the-art would not look significantly different for local LLMs. This is almost a nothingburger if not for the simple novelty of OpenAI releasing an Open AI for once in their life.
- modeless 1y ago> Both of them are comparable to OpenAI's largest models and even bench favorably to Deepseek and Opus So are/do the new OpenAI models, except they're much smaller.
- UrineSqueegee 1y agoI'd really wait for additional neutral benchmarks, I asked the 20b model on low reasoning effort which number is larger 9.9 or 9.11 and it got it wrong. Qwen-0.6b gets it right.
- bigyabai 1y agoAccording to the early benchmarks, it's looking like you're just flat-out wrong: https://blog.brokk.ai/a-first-look-at-gpt-oss-120bs-coding-ability/ https://blog.brokk.ai/a-first-look-at-gpt-oss-120bs-coding-a...