publisher

redhatai

1 build measured across 15 families, and 8 derivatives of other people's models. Ranked by what machines proved, never by downloads.

measured — signed by the machine that ran eachRedHatAI/phi-4-quantized.w8a8phi-4/base · vllm · 32,768 ctx
derivatives — changes this publisher made to someone else's weightsDevstral-Small-2-24B-Instruct-2512devstral-smalldiffusiongemma-26B-A4B-it-NVFP4diffusiongemmadiffusiongemma-26B-A4B-it-FP8-dynamicdiffusiongemmaembeddinggemma-300membeddinggemmaembeddinggemma-2embeddinggemma-2Kimi-K3-speculator.dsparkkimi-k3Mistral-Small-4-119B-2603.dflashmistral-small-4Mistral-Small-4-119B-2603-speculator.dsparkmistral-small-4
in the catalog, not yet measured — declared, unsignedRedHatAI/Devstral-Small-2-24B-Instruct-2512devstral-small/24b-base-2503~redhatai-2-2512 · FP8 · vllm · derivativeRedHatAI/diffusiongemma-26B-A4B-it-NVFP4diffusiongemma/26b-a4b~redhatai · W4A4 · vllm · derivativeRedHatAI/diffusiongemma-26B-A4B-itdiffusiongemma/26b-a4b~redhatai · BF16 · vllm · derivativeRedHatAI/diffusiongemma-26B-A4B-it-FP8-dynamicdiffusiongemma/26b-a4b~redhatai-dynamic · FP8 · vllm · derivativeRedHatAI/embeddinggemma-300membeddinggemma/300m~redhatai-embeddinggemma · FP32 · vllm · derivativeRedHatAI/embeddinggemma-2embeddinggemma-2/base~redhatai · BF16 · vllm · derivativeRedHatAI/gemma-4-26B-A4B-it-FP8-dynamicgemma-4/26b-a4b · FP8 · vllm · quantizerRedHatAI/granite-4.0-h-tiny-FP8-dynamicgranite-4-0/h-tiny · FP8 · vllm · quantizerRedHatAI/Kimi-K2.6-FP8-BLOCKkimi-k2-6/base · FP8 · vllm · quantizerRedHatAI/Kimi-K3-speculator.dsparkkimi-k3/base~redhatai-speculator-dspark · BF16 · vllm · derivativeRedHatAI/Llama-3.2-1B-Instruct-FP8-dynamicllama-3-2/1b · FP8 · vllm · quantizerRedHatAI/Llama-3.2-3B-Instruct-FP8-dynamicllama-3-2/3b · FP8 · vllm · quantizerRedHatAI/Ministral-3-14B-Instruct-2512-FP8-dynamicministral-3/14b-instruct-2512 · FP8 · vllm · quantizerRedHatAI/Mistral-Small-4-119B-2603.dflashmistral-small-4/119b-2603~redhatai-dflash · BF16 · vllm · derivativeRedHatAI/Mistral-Small-4-119B-2603-speculator.dsparkmistral-small-4/119b-2603~redhatai-speculator-dspark · BF16 · vllm · derivativeRedHatAI/Muse-Glimmer-30B-NVFP4muse-glimmer/30b · W4A4 · vllm · quantizerRedHatAI/Muse-Glimmer-30B-FP8-blockmuse-glimmer/30b · FP8 · vllm · quantizerRedHatAI/phi-4-FP8-dynamicphi-4/base · FP8 · vllm · quantizerRedHatAI/Qwen3.6-35B-A3B-NVFP4qwen-3-6/35b-a3b · W4A4 · vllm · quantizerRedHatAI/SmolLM3-3B-FP8-dynamicsmollm-3/3b · FP8 · vllm · quantizer