batiai/Qwen3.8-27B-GGUF
Text Generation β’ 27B β’ Updated β’ 2.83k β’ 1
π₯ Qwen 3.8 β Newest 27B, Korean-verified
Qwen's newest 27B: vision, 262K context, Korean-verified. Start with IQ4_XS β we measured every tier and Q3_K_M is smaller yet slower. Needs 24GB+.
Note Six quants + vision projector. Per-quant Apple Silicon and CUDA speeds in the card β including what we got wrong the first time.
batiai/Qwen3.6-35B-A3B-GGUF
Text Generation β’ 35B β’ Updated β’ 3.4k β’ 5Note MoE alternative β 3B active, 45 t/s on M4 Max vs 16.6 for a dense 27B. The pick for 16GB Macs or when throughput beats peak quality.
batiai/Qwen3.6-27B-GGUF
Text Generation β’ 27B β’ Updated β’ 146k β’ 8Note Previous generation, same class. Our most-downloaded model β a useful baseline for comparison.