4 ms·
remember to let llama.cpp catch up to anything new in this model. Save your judgment until about 2 weeks of use.
by altruios 2mo ago
remember to let llama.cpp catch up to anything new in this model. Save your judgment until about 2 weeks of use.
- chrismartin 2mo ago'Good' news, there seems to be nothing new architecture-wise. Same as Qwen 3.5 and 3.6, so llama.cpp doesn't know the difference.