LocalFit
Submit a benchmark

granite-switch-4.1-8b Composite

ibm-granite/granite-switch-4.1-8b

Released2026-10-14Size8.8B tensors · denseRoleCompositeLicenseapache-2.0Context128K config max

Where it runs 8K context · ≥ 5 t/s floor

SetupBest planMemoryGenerate
8 GB GPU + 32 GB RAMnative I64/F32vLLMGPU streams weights from RAM · 12 GiB in RAMFits≤ 2.5 t/s Ceiling
12 GB GPU + 32 GB RAMnative I64/F32vLLMGPU streams weights from RAM · 8 GiB in RAMFits≤ 3.7 t/s Ceiling
16 GB GPU + 64 GB RAMnative I64/F32vLLMGPU streams weights from RAM · 4 GiB in RAMFits≤ 8.0 t/s Ceiling
24 GB GPU + 64 GB RAMnative I64/F32vLLMFull GPUFits≤ 55 t/s Ceiling
32 GB GPU + 128 GB RAMnative I64/F32vLLMFull GPUFits≤ 98 t/s Ceiling
2× 24 GB GPUs + 128 GB RAMnative I64/F32vLLMFull GPUFits≤ 51 t/s Ceiling
96 GB workstation GPU + 256 GB RAMnative I64/F32vLLMFull GPUFits≤ 98 t/s Ceiling
Mac, 24 GBDoes not fit
Mac, 48 GBDoes not fit
Mac, 128 GBDoes not fit
Mac, 512 GBDoes not fit
DGX Spark 128 GBnative I64/F32vLLMUnified memoryFits≤ 15 t/s Ceiling
Ryzen AI Max+ 395, 128 GBnative I64/F32vLLMUnified memoryFits≤ 14 t/s Ceiling
CPU only, 64 GB DDR5-5600 dual channelDoes not fit

Same plan engine as the planner, with its default reserves and RAM. How plans are chosen.

Files exact Hugging Face bytes

ArtifactSizeFilesRevisionSource
Native safetensors mixed/custom native; ; stored tensor dtypes=I64/F32/BF16 · I
ibm-granite/granite-switch-4.1-8b
16.35 GiB40d1625d16dHF files

Not found in the representative quant search (not proven unavailable): GGUF FP16, GGUF Q8_0, GGUF Q5_K_M, GGUF Q4_K_M, MLX 8-bit, MLX 4-bit, EXL3, AWQ, GPTQ. Complete selected file sets; projectors, MTP/draft and lookup files are listed separately and budgeted only by plans that use them. File bytes are not allocated memory.

Quality

No Artificial Analysis evaluation mapped to this checkpoint.

Arena text rating

Unknown

Publisher-reported benchmarks

None extracted from the model card.

AA and Arena describe the hosted model in the stated mode, not each quantized file; scores are never inherited by base, pruned, abliterated or quantized variants and never averaged together.

How to run it

vLLM withgranite-switch · unknown

publisher-documented, not executed source

git clone https://github.com/generative-computing/granite-switch.git
cd granite-switch

# Pick what you need:
pip install -e ".[compose]"   # Compose models with adapters
pip install -e ".[hf]"        # HuggingFace inference
pip install -e ".[vllm]"      # vLLM inference
pip install -e ".[audio]"     # Audio input (adds vLLM's audio deps)
pip install -e ".[dev]"       # Everything
pip install -e ".[vllm]"

python -m vllm.entrypoints.openai.api_server \
  --model ibm-granite/granite-switch-4.1-8b \
  --port 8000

Prerequisites: See source setup/dependencies; these snippets were inspected, not executed.

  • Compose dLLM +adapters+speech encoder, not standalone base weight count.
  • Activation/control tokens and adapter loading essential; base LLM alone does not implement decisions or speech.
  • Speech encoder is separate ASR cascade, not jointly trained native audio.

Recipes are source examples (settings are the publisher’s or author’s, not measured optima) and were not executed by LocalFit. A downloadable file does not prove your backend build supports the architecture.

Reported speeds 0

No reported measurement yet. Submit one.