3 ms·
> Models that you can run at home (Like Qwen 35B) aren't remotely close to Opus or GPT 5.5. Is that characterization based on some objective facts or benchmark
by pbasista 4mo ago
> Models that you can run at home (Like Qwen 35B) aren't remotely close to Opus or GPT 5.5.
Is that characterization based on some objective facts or benchmarks?
- redox99 4mo agoBased on private test prompts I've run through OpenRouter.
- kube-system 4mo agoYes, there aren't any 35B models that are beating frontier models at just about anything generalized