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-128gb. Full machine surface: /llms.txt · REST API: api.bestmodel.run rig | model | basis | value | n gb10-grace-blackwell-128gb | nvidia-gemma-4-26b-a4b-nvfp4 | reported | 130.5 tok/s | n=2 gb10-grace-blackwell-128gb | nvidia-qwen3-6-35b-a3b-nvfp4 | measured | 104.9 tok/s | n=4 gb10-grace-blackwell-128gb | redhatai-qwen3-6-35b-a3b-nvfp4 | measured | 94.8 tok/s | n=3 gb10-grace-blackwell-128gb | aeon-7-qwen3-6-35b-a3b-heretic-nvfp4 | measured | 82.3 tok/s | n=4 gb10-grace-blackwell-128gb | poolside-laguna-s-2-1-nvfp4 | measured | 73.8 tok/s | n=4 gb10-grace-blackwell-128gb | bartowski-kwaipilot-kat-coder-v2-5-dev-gguf | reported | 72 tok/s | n=1 gb10-grace-blackwell-128gb | nvidia-nemotron-3-nano-omni-30b-a3b-reasoning-nvfp4 | reported | 71.6 tok/s | n=1 gb10-grace-blackwell-128gb | unsloth-qwen3-6-35b-a3b-nvfp4-fast | reported | 70.6 tok/s | n=1 gb10-grace-blackwell-128gb | qwen-qwen3-6-35b-a3b | reported | 64.5 tok/s | n=2 gb10-grace-blackwell-128gb | z-lab-qwen3-6-27b-dflash | reported | 62.1 tok/s | n=1 gb10-grace-blackwell-128gb | unsloth-qwen3-5-35b-a3b-gguf | reported | 60.3 tok/s | n=1 gb10-grace-blackwell-128gb | qwen-qwen3-6-35b-a3b | reported | 56.1 tok/s | n=1 gb10-grace-blackwell-128gb | qwen-qwen3-6-35b-a3b | measured | 55.7 tok/s | n=3 gb10-grace-blackwell-128gb | qwen-qwen3-5-35b-a3b-fp8 | reported | 51.4 tok/s | n=1 gb10-grace-blackwell-128gb | unsloth-gemma-4-26b-a4b-it-gguf | reported | 50.7 tok/s | n=1 gb10-grace-blackwell-128gb | openai-gpt-oss-20b | reported | 49.2 tok/s | n=1 gb10-grace-blackwell-128gb | qwen-qwen-agentworld-35b-a3b | reported | 49.1 tok/s | n=2 gb10-grace-blackwell-128gb | deepseek-ai-deepseek-v4-flash-dspark | reported | 46.4 tok/s | n=1 gb10-grace-blackwell-128gb | deepseek-ai-deepseek-v4-flash | reported | 45.7 tok/s | n=1 gb10-grace-blackwell-128gb | qwen-qwen3-5-122b-a10b | reported | 45.6 tok/s | n=1 gb10-grace-blackwell-128gb | lmstudio-community-olmocr-2-7b-1025-gguf | reported | 41.6 tok/s | n=1 gb10-grace-blackwell-128gb | radixark-qwen3-8-flash-next-nvfp4 | reported | 40.9 tok/s | n=1 gb10-grace-blackwell-128gb | quanttrio-qwen3-6-35b-a3b-awq | reported | 39.7 tok/s | n=1 gb10-grace-blackwell-128gb | nvidia-qwen3-next-80b-a3b-instruct-nvfp4 | reported | 39 tok/s | n=1 gb10-grace-blackwell-128gb | unsloth-gemma-4-26b-a4b-it-gguf | reported | 38.8 tok/s | n=1 gb10-grace-blackwell-128gb | nvidia-nvidia-nemotron-labs-3-puzzle-75b-a9b-nvfp4 | reported | 38.7 tok/s | n=1 gb10-grace-blackwell-128gb | orcarouter-qwen3-8-27b-uncensored-nvfp4 | reported | 37.7 tok/s | n=1 gb10-grace-blackwell-128gb | inclusionai-ling-3-0-flash-int4 | reported | 37.3 tok/s | n=1 gb10-grace-blackwell-128gb | unsloth-qwen3-5-35b-a3b-gguf | reported | 35.1 tok/s | n=1 gb10-grace-blackwell-128gb | openai-gpt-oss-120b | reported | 34.4 tok/s | n=1 gb10-grace-blackwell-128gb | aeon-7-qwen3-6-27b-aeon-ultimate-uncensored-nvfp4 | measured | 33.5 tok/s | n=4 gb10-grace-blackwell-128gb | qwen-qwen3-6-27b | measured | 33.3 tok/s | n=6 gb10-grace-blackwell-128gb | sakamakismile-qwen3-6-27b-nvfp4 | reported | 32.5 tok/s | n=2 gb10-grace-blackwell-128gb | intel-qwen3-5-122b-a10b-int4-autoround | reported | 31 tok/s | n=2 gb10-grace-blackwell-128gb | unsloth-minimax-m2-7 | reported | 30.9 tok/s | n=2 gb10-grace-blackwell-128gb | minimaxai-minimax-m2-7 | reported | 30.9 tok/s | n=1 gb10-grace-blackwell-128gb | aeon-7-qwen3-6-27b-aeon-ultimate-uncensored-multimodal-nvfp4-mtp-xs | measured | 29.8 tok/s | n=7 gb10-grace-blackwell-128gb | unsloth-qwen3-8-flash-next-gguf | reported | 27.3 tok/s | n=1 gb10-grace-blackwell-128gb | lukealonso-mimo-v2-5-nvfp4 | measured | 27.2 tok/s | n=3 gb10-grace-blackwell-128gb | cyankiwi-minimax-m2-7-awq-4bit | measured | 27 tok/s | n=4 gb10-grace-blackwell-128gb | zai-org-glm-4-7-flash | reported | 27 tok/s | n=2 gb10-grace-blackwell-128gb | unsloth-qwen3-6-27b-nvfp4 | measured | 26.4 tok/s | n=3 gb10-grace-blackwell-128gb | google-gemma-4-12b | reported | 25.8 tok/s | n=1 gb10-grace-blackwell-128gb | radixark-qwen3-8-27b-nvfp4-bf16-lmhead | measured | 24.6 tok/s | n=4 gb10-grace-blackwell-128gb | sakamakismile-huihui-qwen3-6-27b-abliterated-nvfp4-mtp | reported | 24 tok/s | n=1 gb10-grace-blackwell-128gb | radixark-qwen3-8-27b-nvfp4 | measured | 23.6 tok/s | n=5 gb10-grace-blackwell-128gb | zai-org-glm-5-3-flash | reported | 23.5 tok/s | n=1 gb10-grace-blackwell-128gb | 0xsero-deepseek-v4-flash-180b | reported | 23.3 tok/s | n=1 gb10-grace-blackwell-128gb | stepfun-ai-step-3-7-flash-gguf | reported | 23 tok/s | n=1 gb10-grace-blackwell-128gb | poolside-laguna-s-2-1-gguf | reported | 19.3 tok/s | n=1 gb10-grace-blackwell-128gb | unsloth-qwen3-8-27b-nvfp4 | measured | 19.1 tok/s | n=6 gb10-grace-blackwell-128gb | jiunsong-superhy3-abliterated-nvfp4 | reported | 19 tok/s | n=1 gb10-grace-blackwell-128gb | rdtand-qwen3-6-27b-prismascout-blackwell-nvfp4-bf16-vllm | reported | 18.9 tok/s | n=1 gb10-grace-blackwell-128gb | qwen-qwen3-5-122b-a10b | reported | 17.1 tok/s | n=1 gb10-grace-blackwell-128gb | qwen-qwen3-235b-a22b-thinking-2507 | reported | 16 tok/s | n=1 gb10-grace-blackwell-128gb | qwen-qwen3-8-27b-fp8 | reported | 14.8 tok/s | n=1 gb10-grace-blackwell-128gb | empero-ai-qwythos-9b-claude-mythos-5-1m | reported | 12.9 tok/s | n=1 gb10-grace-blackwell-128gb | qwen-qwen3-5-27b | measured | 11.5 tok/s | n=4 gb10-grace-blackwell-128gb | unsloth-qwen3-5-27b-gguf | reported | 11.1 tok/s | n=1 gb10-grace-blackwell-128gb | antirez-deepseek-v4-gguf | reported | 10.5 tok/s | n=1 Capture or correct a number: /console (signed, Ed25519).