B300

22 records across 7 models (quants: BF16, FP4, FP8, MXFP4, NVFP4, as-published).

hardwaremodelquantbackendtok/spp tok/sWttft sctx/tgdateclassflagssource
B300DeepSeek-V4-Flash MoEFP4vLLM1216––––2026-07community–source
B300DeepSeek-V4-Flash MoEFP4vLLM773––––2026-07community–source
B300DeepSeek-V4-Flash MoEFP4vLLM344––––2026-07community–source
B300DiffusionGemma-26B-A4B BF16vLLM1152––0.39–2026-10sourced–source
B300DiffusionGemma-26B-A4B BF16vLLM947––0.55–2026-10sourced–source
B300GLM-5.3 MXFP4vLLM91.6–––10242026-10-22sourced–source
B300GLM-5.3 MXFP4vLLM86.66–––10242026-10-22sourced–source
B300Llama 3.3 70B 70B denseNVFP4vLLM162.29––––2026-06community–source
B300Qwen3 4B BF16vLLM10543–––tg2562026-10-02community–source
B300Qwen3 4B BF16vLLM10495–––tg2562026-10-02community–source
B300Qwen3 4B BF16vLLM3032–––tg2562026-10-02community–source
B300Qwen3 4B BF16vLLM3012–––tg2562026-10-02community–source
B300Qwen3 4B BF16vLLM464.7–––tg2562026-10-02community–source
B300Qwen3 4B BF16vLLM462.3–––tg2562026-10-02community–source
B300Qwen3.6-35B-A3B as-publishedvLLM733––––2026-09-25community–source
B300Qwen3.6-35B-A3B as-publishedvLLM460––––2026-09-25community–source
B300Qwen3.6-35B-A3B-FP8 35BFP8vLLM7221.04––––2026-09-29community–source
B300Qwen3.6-35B-A3B-FP8 35BFP8vLLM5115.53––––2026-09-29community–source
B300Qwen3.6-35B-A3B-FP8 35BFP8vLLM427.01––––2026-09-29community–source
B300Qwen3.6-35B-A3B-FP8 35BFP8vLLM286.56––––2026-09-29community–source
B300Qwen3.6-35B-A3B-FP8 35BFP8vLLM7.015––––2026-09-29community–source
B300Qwen3.6-35B-A3B-FP8 35BFP8vLLM6.419––––2026-09-29community–source

All rows cite their source. Provenance: community 40018 sourced 1085

← all hardware