Granite 4.0 H Small — Hardware Requirements & GPU Compatibility
Chat
Granite-4.0-H-Small is IBM's 32-billion-parameter instruction-tuned model in the Granite 4.0 line, fine-tuned from Granite-4.0-H-Small-Base for enterprise use: summarization, classification, extraction, question answering, retrieval-augmented generation, code tasks, function calling, and multilingual dialog. Architecturally it is a hybrid mixture-of-experts model with only 4 full-attention layers against 36 Mamba-2 state-space layers, and 72 experts with 10 active per token, giving about 9 billion active parameters out of 32 billion total; the Mamba-2 layers avoid growing a KV cache with context length, keeping long-context inference cheaper than a pure attention stack. It was trained with supervised fine-tuning, RL-based alignment, and model merging. At 32 billion parameters, it needs a high-end consumer GPU or multi-GPU setup once quantized. Context length is 131,072 tokens (128K). It is released under the Apache 2.0 license, permitting unrestricted commercial and research use, and was published in October 2025.
17.7K downloads
309 likesSpecifications
- Publisher
- IBM
- Family
- Granite
- Parameters
- 32.2B
- Architecture
- GraniteMoeHybridForCausalLM
- Context Length
- 131,072 tokens
- Vocabulary Size
- 100,352
- Release Date
- 2025-09-16
- License
- Apache 2.0
Get Started
HuggingFace
How Much VRAM Does Granite 4.0 H Small Need?
Select a quantization to see compatible GPUs below.
| Quantization | Bits | VRAM | + Context | File Size | Quality |
|---|---|---|---|---|---|
| Q2_K | 3.40 | 14.3 GB | 35.5 GB | 13.69 GB | 2-bit quantization with K-quant improvements |
| Q3_K_S | 3.50 | 14.7 GB | 35.9 GB | 14.09 GB | 3-bit small quantization |
| Q3_K_M | 3.90 | 16.3 GB | 37.5 GB | 15.70 GB | 3-bit medium quantization |
| Q4_0 | 4.00 | 16.7 GB | 37.9 GB | 16.10 GB | 4-bit legacy quantization |
| Q4_K_M | 4.80 | 20.0 GB | 41.1 GB | 19.32 GB | 4-bit medium quantization — most popular sweet spot |
| Q5_K_M | 5.70 | 23.6 GB | 44.7 GB | 22.95 GB | 5-bit medium quantization — good quality/size tradeoff |
| Q6_K | 6.60 | 27.2 GB | 48.4 GB | 26.57 GB | 6-bit quantization, very good quality |
| Q8_0 | 8.00 | 32.8 GB | 54.0 GB | 32.21 GB | 8-bit quantization, near-lossless |
Which GPUs Can Run Granite 4.0 H Small?
Q4_K_M · 20.0 GBGranite 4.0 H Small (Q4_K_M) requires 20.0 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 26+ GB is recommended. Using the full 131K context window can add up to 21.1 GB, bringing total usage to 41.1 GB. 8 GPUs can run it, including NVIDIA GeForce RTX 5090, NVIDIA GeForce RTX 3090 Ti.
Runs great
— Plenty of headroom
Which Devices Can Run Granite 4.0 H Small?
Q4_K_M · 20.0 GB41 devices with unified memory can run Granite 4.0 H Small, including NVIDIA DGX H100, NVIDIA DGX A100 640GB, Mac Mini M4 Pro (24 GB).
Runs great
— Plenty of headroomDecent
— Enough memory, may be tightWhere to Download Granite 4.0 H Small
Community quantizations of this model — GGUF for llama.cpp, Ollama, and LM Studio, plus AWQ/MLX variants where available.
Benchmarks
Benchmark details →Related Models
Frequently Asked Questions
- How much VRAM does Granite 4.0 H Small need?
Granite 4.0 H Small requires 20.0 GB of VRAM at Q4_K_M, or 65.0 GB at BF16. Full 131K context adds up to 21.1 GB (41.1 GB total).
VRAM = Weights + KV Cache + Overhead
Weights = 32.2B × 4.8 bits ÷ 8 = 19.3 GB
KV Cache + Overhead ≈ 0.7 GB (at 2K context + ~0.3 GB framework)
KV Cache + Overhead ≈ 21.8 GB (at full 131K context)
VRAM usage by quantization
Q4_K_M20.0 GBQ4_K_M + full context41.1 GB- Can NVIDIA GeForce RTX 4090 run Granite 4.0 H Small?
Yes, at Q5_K_L (24.0 GB) or lower. Higher quantizations like Q6_K (27.2 GB) exceed the NVIDIA GeForce RTX 4090's 24 GB.
- What's the best quantization for Granite 4.0 H Small?
For Granite 4.0 H Small, Q4_K_M (20.0 GB) offers the best balance of quality and VRAM usage. Q4_K_L (20.4 GB) provides better quality if you have the VRAM. The smallest option is IQ2_XXS at 9.5 GB.
VRAM requirement by quantization
IQ2_XXS9.5 GBQ2_K14.3 GBQ4_118.8 GBQ4_K_M ★20.0 GBQ5_122.8 GBBF1665.0 GB★ Recommended — best balance of quality and VRAM usage.
- Can I run Granite 4.0 H Small on a Mac?
Granite 4.0 H Small requires at least 9.5 GB at IQ2_XXS, which exceeds the unified memory of most consumer Macs. You would need a Mac Studio or Mac Pro with a high-memory configuration.
- Can I run Granite 4.0 H Small locally?
Yes — Granite 4.0 H Small can run locally on consumer hardware. At Q4_K_M quantization it needs 20.0 GB of VRAM. Popular tools include Ollama, LM Studio, and llama.cpp.
- How fast is Granite 4.0 H Small?
At Q4_K_M, Granite 4.0 H Small can reach ~110 tok/s on AMD Instinct MI350X. On NVIDIA GeForce RTX 4090: ~95 tok/s. Speed depends mainly on GPU memory bandwidth. Real-world results typically within ±20%.
tok/s = (bandwidth GB/s ÷ model GB) × efficiency
Example: NVIDIA B200 → 8000 ÷ 20.0 × 0.65 = ~308 tok/s
Estimated speed at Q4_K_M (20.0 GB)
~308 tok/s~95 tok/s~308 tok/s~256 tok/sReal-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
- What's the download size of Granite 4.0 H Small?
At Q4_K_M, the download is about 19.32 GB. The full-precision BF16 version is 64.41 GB. The smallest option (IQ2_XXS) is 8.86 GB.
- Which GPUs can run Granite 4.0 H Small?
8 consumer GPUs can run Granite 4.0 H Small at Q4_K_M (20.0 GB). Top options include NVIDIA GeForce RTX 5090, AMD Radeon RX 7900 XT, AMD Radeon RX 7900 XTX. 1 GPU have plenty of headroom for comfortable inference.
- Which devices can run Granite 4.0 H Small?
41 devices with unified memory can run Granite 4.0 H Small at Q4_K_M (20.0 GB), including AMD Ryzen AI 9 HX 370 (Strix Point) Laptop, ASUS Ascent GX10, Asus ROG Flow Z13 (2025, Ryzen AI Max+ 395, 128 GB), Beelink GTR9 Pro (Ryzen AI Max+ 395, 128 GB). Apple Silicon Macs use unified memory shared between CPU and GPU, making them well-suited for local LLM inference.