#vram

3 posts

1. 2026-08-19
   Dense 27B @168k token prompt

   • #long-context
   • #vram
   • #ollama
2. 2026-07-06
   Max context in 48 GB: five models

   • #vram
   • #ollama
   • #gemma
   • #qwen
3. 2026-07-06
   Qwen3.6 35B-A3B MoE vs 27B @26k token prompt

   • #moe
   • #long-context
   • #vram
   • #ollama