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=radeon-890m-32gb. Full machine surface: /llms.txt · REST API: api.bestmodel.run rig | model | basis | value | n radeon-890m-32gb | unsloth-gemma-4-e2b-it-gguf | reported | 35.2 tok/s | n=1 radeon-890m-32gb | microsoft-phi-4-mini-reasoning | reported | 33.2 tok/s | n=1 radeon-890m-32gb | unsloth-qwen3-30b-a3b-gguf | reported | 30.7 tok/s | n=1 radeon-890m-32gb | unsloth-qwen3-coder-30b-a3b-instruct-gguf | reported | 30 tok/s | n=1 radeon-890m-32gb | google-gemma-4-e4b-it | reported | 26.7 tok/s | n=1 radeon-890m-32gb | openai-gpt-oss-20b | reported | 23.9 tok/s | n=1 radeon-890m-32gb | bartowski-nvidia-nemotron-cascade-2-30b-a3b-gguf | reported | 21.5 tok/s | n=1 radeon-890m-32gb | unsloth-qwen3-6-35b-a3b-gguf | reported | 20.1 tok/s | n=1 radeon-890m-32gb | unsloth-gemma-4-e4b-it-gguf | reported | 18.7 tok/s | n=1 radeon-890m-32gb | qwen-qwen3-8b | reported | 15.9 tok/s | n=2 radeon-890m-32gb | lmstudio-community-olmocr-2-7b-1025-gguf | reported | 14.9 tok/s | n=1 radeon-890m-32gb | unsloth-qwen3-vl-8b-instruct-gguf | reported | 14.1 tok/s | n=1 radeon-890m-32gb | google-gemma-4-12b-it | reported | 10.4 tok/s | n=1 radeon-890m-32gb | google-gemma-4-12b-it | reported | 8.5 tok/s | n=1 radeon-890m-32gb | google-gemma-4-12b-it | reported | 6.8 tok/s | n=1 radeon-890m-32gb | qwen-qwen3-6-27b | reported | 4.9 tok/s | n=1 radeon-890m-32gb | unsloth-qwen3-5-27b-gguf | reported | 4.2 tok/s | n=1 Capture or correct a number: /console (signed, Ed25519).