bestmodel.run / pool — agent view

Honesty ladder: measured > reported > extrapolated > formula > no data yet.
Top 60 cells, default sort: fastest decode. Columns: rig | model | basis | value | n.
Filter: rig=rtx-4090-24gb.
Full machine surface: /llms.txt · REST API: api.bestmodel.run

  rig                               | model                                  | basis    | value          | n
  rtx-4090-24gb                      | unsloth-qwen3-6-35b-a3b-mtp-gguf       | reported | 213.9 tok/s    | n=1
  rtx-4090-24gb                      | unsloth-qwen3-6-35b-a3b-gguf           | reported | 165.1 tok/s    | n=1
  rtx-4090-24gb                      | llmfan46-qwen3-6-35b-a3b-uncensored-heretic | reported | 164.1 tok/s    | n=1
  rtx-4090-24gb                      | qwen-qwen3-6-35b-a3b                   | reported | 157.7 tok/s    | n=2
  rtx-4090-24gb                      | unsloth-qwen3-5-35b-a3b-gguf           | reported | 157.4 tok/s    | n=1
  rtx-4090-24gb                      | unsloth-qwen3-6-27b-mtp-gguf           | reported | 94.1 tok/s     | n=2
  rtx-4090-24gb                      | liconstudio-gemma-4-31b-it-abliterated-gguf | reported | 87.7 tok/s     | n=1
  rtx-4090-24gb                      | lorbus-qwen3-6-27b-int4-autoround      | reported | 52.6 tok/s     | n=2
  rtx-4090-24gb                      | unsloth-qwen3-6-27b                    | reported | 49.4 tok/s     | n=1
  rtx-4090-24gb                      | unsloth-qwen3-5-27b-gguf               | reported | 45.4 tok/s     | n=2
  rtx-4090-24gb                      | qwen-qwen3-6-27b                       | measured | 44.2 tok/s     | n=3
  rtx-4090-24gb                      | unsloth-qwen3-6-27b-gguf               | reported | 41.8 tok/s     | n=1

Capture or correct a number: /console (signed, Ed25519).