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=arc-pro-b70-32gb-x2.
Full machine surface: /llms.txt · REST API: api.bestmodel.run

  rig                               | model                                  | basis    | value          | n
  arc-pro-b70-32gb-x2                | redhatai-qwen3-5-4b-quantized-w4a16    | reported | 236.9 tok/s    | n=1
  arc-pro-b70-32gb-x2                | redhatai-qwen3-5-9b-quantized-w4a16    | reported | 172.3 tok/s    | n=1
  arc-pro-b70-32gb-x2                | redhatai-qwen3-5-9b-fp8-dynamic        | reported | 147.8 tok/s    | n=1
  arc-pro-b70-32gb-x2                | qwen-qwen3-6-35b-a3b                   | reported | 110.7 tok/s    | n=1
  arc-pro-b70-32gb-x2                | devan-carlin-qwen3-8-27b-int4-autoround | measured | 107.1 tok/s    | n=4
  arc-pro-b70-32gb-x2                | qwen-qwen3-coder-30b-a3b-instruct      | reported | 93.3 tok/s     | n=1
  arc-pro-b70-32gb-x2                | webhie-qwen3-6-27b-int4-autoround      | measured | 89.4 tok/s     | n=6
  arc-pro-b70-32gb-x2                | nameistoken-qwen3-6-35b-a3b-quark-w8a8-int8 | reported | 85.9 tok/s     | n=1
  arc-pro-b70-32gb-x2                | qwen-qwen3-6-35b-a3b                   | reported | 73.9 tok/s     | n=1
  arc-pro-b70-32gb-x2                | goldhub-qwen3-8-27b-int4-w4a16-autoround | reported | 67.7 tok/s     | n=1
  arc-pro-b70-32gb-x2                | qwen-qwen3-8-27b-fp8                   | measured | 64.3 tok/s     | n=12
  arc-pro-b70-32gb-x2                | openai-gpt-oss-120b                    | reported | 59.4 tok/s     | n=1
  arc-pro-b70-32gb-x2                | qwen-qwen3-coder-next                  | reported | 58.5 tok/s     | n=1
  arc-pro-b70-32gb-x2                | lorbus-qwen3-6-27b-int4-autoround      | measured | 48.3 tok/s     | n=3
  arc-pro-b70-32gb-x2                | ggml-org-qwen3-8-27b-gguf              | measured | 45.9 tok/s     | n=3
  arc-pro-b70-32gb-x2                | unsloth-qwen3-6-27b                    | reported | 42.1 tok/s     | n=1
  arc-pro-b70-32gb-x2                | qwen-qwen3-6-27b                       | reported | 40.5 tok/s     | n=1
  arc-pro-b70-32gb-x2                | nvidia-nvidia-nemotron-labs-3-puzzle-75b-a9b-bf16 | reported | 39.1 tok/s     | n=1
  arc-pro-b70-32gb-x2                | z-lab-qwen3-6-27b-dflash               | measured | 38.2 tok/s     | n=4
  arc-pro-b70-32gb-x2                | atomicchat-qwen3-8-flash-next-gguf     | measured | 27.7 tok/s     | n=3
  arc-pro-b70-32gb-x2                | qwen-qwen3-6-27b                       | reported | 25.7 tok/s     | n=1
  arc-pro-b70-32gb-x2                | qwen-qwen3-8-flash-next                | reported | 24.5 tok/s     | n=1
  arc-pro-b70-32gb-x2                | qwen-qwen3-8-flash-next                | reported | 21.9 tok/s     | n=2
  arc-pro-b70-32gb-x2                | qwen-qwen3-6-27b-fp8                   | reported | 20.1 tok/s     | n=1

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