GPUfit — NVIDIA B200https://gpufit.tech/gpu/b200New catalog models scored against NVIDIA B200.bottlecapai ThinkingCap gemma 4 26B A4B it on NVIDIA B200: Runs greathttps://gpufit.tech/models/bottlecapai-thinkingcap-gemma-4-26b-a4b-ithttps://gpufit.tech/models/bottlecapai-thinkingcap-gemma-4-26b-a4b-itFri, 23 Oct 2026 21:26:38 GMTRuns great fully on your NVIDIA B200 at BF16, 8K context, about 439–599 tok/s. Weights 47.0 GB plus 0.51 GB of KV cache — 48.3 GB of 182.4 GB usable. Up to 256K context still fits. 439–599 tok/s.NVIDIA Nemotron 3.5 Bolt E4B on NVIDIA B200: Runs greathttps://gpufit.tech/models/nvidia-nemotron-3-5-bolt-e4bhttps://gpufit.tech/models/nvidia-nemotron-3-5-bolt-e4bFri, 23 Oct 2026 09:47:55 GMTRuns great fully on your NVIDIA B200 at BF16, 8K context, about 279–381 tok/s. Weights 14.0 GB plus 0.14 GB of KV cache — 14.6 GB of 182.4 GB usable. Up to 128K context still fits. 279–381 tok/s.DeepSeek V4.1 Flash 1017 on NVIDIA B200: Won't fithttps://gpufit.tech/models/deepseek-v4-1-flash-1017https://gpufit.tech/models/deepseek-v4-1-flash-1017Sat, 17 Oct 2026 20:36:06 GMTWon't fit on NVIDIA B200. The smallest quant UD-IQ1_S is 217.0 GB before KV cache, and usable memory is 182.4 GB.LFM2.5 24B A2B on NVIDIA B200: Runs greathttps://gpufit.tech/models/lfm2-5-24b-a2bhttps://gpufit.tech/models/lfm2-5-24b-a2bTue, 13 Oct 2026 14:19:48 GMTRuns great fully on your NVIDIA B200 at BF16, 8K context, about 942–1284 tok/s. Weights 44.7 GB plus 0.63 GB of KV cache — 46.0 GB of 182.4 GB usable. Up to 128K context still fits. 942–1284 tok/s.