@IPngNetworks @jahanson Yep, Claude is... hmm. I prefer local models like Llama4-Scout, Qwen3 (huge context window), Gemma3, and IBM's GraniteCode 3.
some of those can be run for free on the Cerebras cloud cluster: https://inference.cerebras.ai
- https://www.ibm.com/granite/docs/models/code/ (inc how to run local models)
- https://huggingface.co/collections/ibm-granite/granite-code-models-6624c5cec322e4c148c8b330
- https://huggingface.co/collections/google/gemma-3-release-67c6c6f89c4f76621268bb6d
- https://huggingface.co/collections/Qwen/qwen3-67dd247413f0e2e4f653967f
- https://huggingface.co/collections/meta-llama/llama-4-67f0c30d9fe03840bc9d0164