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=gb10-grace-blackwell-256gb. Full machine surface: /llms.txt · REST API: api.bestmodel.run rig | model | basis | value | n gb10-grace-blackwell-256gb | fraserprice-deepseek-v4-flash-dspark | reported | 65.1 tok/s | n=1 gb10-grace-blackwell-256gb | nvidia-qwen3-8-flash-next-nvfp4 | reported | 62.9 tok/s | n=1 gb10-grace-blackwell-256gb | drowzeys-keys-deepseekv4-flash-ga-0731-dspark-abliterated-32-32 | reported | 39.8 tok/s | n=1 gb10-grace-blackwell-256gb | jiunsong-superdeepseek-v4-flash-abliterated-mq-2xdgx | reported | 37.3 tok/s | n=1 gb10-grace-blackwell-256gb | brandonmusic-glm-5-3-flash-tr3-4bpw | reported | 36.6 tok/s | n=1 gb10-grace-blackwell-256gb | jiunsong-superdeepseek-v4-flash-abliterated-mq-2xdgx | reported | 36.3 tok/s | n=1 gb10-grace-blackwell-256gb | deepseek-ai-deepseek-v4-flash | reported | 36 tok/s | n=1 gb10-grace-blackwell-256gb | zai-org-glm-5-3-flash | measured | 32.6 tok/s | n=8 gb10-grace-blackwell-256gb | deepseek-ai-deepseek-v4-flash | reported | 26.9 tok/s | n=2 gb10-grace-blackwell-256gb | deepseek-ai-deepseek-v4-flash-0731 | measured | 24.9 tok/s | n=3 gb10-grace-blackwell-256gb | lukealonso-minimax-m2-5-nvfp4 | reported | 17.7 tok/s | n=1 gb10-grace-blackwell-256gb | unsloth-qwen3-8-27b-nvfp4 | reported | 17.4 tok/s | n=1 gb10-grace-blackwell-256gb | pastapaul-deepseek-v4-flash-w4a16-fp8 | reported | 11.9 tok/s | n=1 gb10-grace-blackwell-256gb | qwen-qwen3-8-27b | reported | 8.2 tok/s | n=1 gb10-grace-blackwell-256gb | 0xsero-glm-5-2-reap-504b-gguf | reported | 7.8 tok/s | n=1 Capture or correct a number: /console (signed, Ed25519).