Is GLM4.7-Flash really the best agentic local llm ? Benchmarks

Z.ai GLM4.7-Flash 30B A3B is a great alternative to gpt-oss 20B for coding and agentinc use cases. It run 100% offline with llamacpp/ollama but currently it thinks a lot and the first quantized release wasn't optimal. https://huggingface.co/unsloth/GLM-4.7-Flash-GGUF https://artificialanalysis.ai/evaluations/tau2-bench Create you personal Local LLM benchmark https://github.com/grigio/llm-eval-simple 👍 Like and subscribe to my channel https://www.youtube.com/channel/UCnVlTfUkoty16uEAsUs6_7g