Granite 3.2 8B Instruct
By IBM · United States
Updated 2026-07-13
Overview
IBM's enterprise-focused 8B Granite 3.2 with a toggleable thinking mode under Apache 2.0. MMLU 65.5 and IFEval 70.9, with built-in IBM safety guardrails.
When to pick this model
- Enterprise RAG deployments needing strict instruction following
- Regulated environments requiring safety guardrails out of the box
- Internal tools where Apache 2.0 plus IBM backing matters
- Workloads benefiting from optional thinking mode
VRAM requirements by quantization
| Quantization | VRAM required |
|---|---|
| Q4_K_M (recommended) | 5 GB |
| Q5_K_M | 6 GB |
| Q8_0 | 9 GB |
| FP16 (no quantization) | 16 GB |
VRAM figures include model weights plus a typical 8k KV cache and ~600 MB runtime overhead (Ollama / llama.cpp baseline). Add headroom for higher context lengths.
In practice, Granite 3.2 8B Instruct fits an 8 GB consumer card at Q4_K_M (5 GB). Stepping up to Q8_0 nearly doubles the footprint to 9 GB, and unquantized FP16 weights take 16 GB — plan your GPU around the Q4 or Q5 figure unless you specifically need the higher fidelity.
Without a GPU, Granite 3.2 8B Instruct needs roughly 10 GB of system RAM to run on CPU via llama.cpp or Ollama — workable for background jobs, but far slower than GPU inference. Throughput estimates from our compatibility engine: around 10 tokens/sec on entry-level GPUs, on the order of 30 tokens/sec on a mid-range card, and up to 80 tokens/sec on high-end hardware — assuming the chosen quantization fully fits in VRAM.
What hardware do you need
The table below matches Granite 3.2 8B Instruct to common GPU memory tiers, using the highest-fidelity quantization that fully fits each card class. Spilling layers to system RAM works but costs most of the speed, so size your card to the quantization you actually want to run.
| GPU memory | Example cards | Best fit for Granite 3.2 8B Instruct |
|---|---|---|
| 8 GB | RTX 5070 Laptop, RTX 5060, RTX 5060 Ti 8GB | Q5_K_M (6 GB used) |
| 12 GB | RTX 5070, RTX 5070 Ti Laptop, RTX 4080 Laptop | Q8_0 (9 GB used) |
| 16 GB | RTX 5080, RTX 4080 Super, Radeon RX 9070 XT | FP16 (16 GB used) |
| 24 GB | RTX 4090, Radeon RX 7900 XTX, RTX 5090 Laptop | FP16 (16 GB used) |
| 32 GB | RTX 5090 | FP16 (16 GB used) |
Which GPU should you buy to run Granite 3.2 8B Instruct?
To run Granite 3.2 8B Instruct locally at Q4, you need ~5 GB of VRAM. The best value for this is a RTX 5060 (8 GB VRAM).
As an Amazon Associate, BestLLMfor earns from qualifying purchases, at no extra cost to you. It does not influence our independent rankings.
Published benchmark scores
| Benchmark | Score |
|---|---|
| MMLU | 67 |
| HumanEval | 72 |
Scores published by the model author or aggregated from public leaderboards. Re-measured monthly by our editorial team.
To put Granite 3.2 8B Instruct in context: its MMLU score of 67 ranks #27 of the 34 catalog models with a published MMLU result (catalog median 73.4); its HumanEval score of 72 ranks #19 of the 26 catalog models with a published HumanEval result (catalog median 81.1). Scores are author-published and measured under different harnesses, so treat gaps of a few points as noise rather than a verdict.
Strengths
- 128k context
- Apache 2.0 license
- Strong RAG and enterprise instruction following
- IBM Safety Guardrails included
- Toggleable thinking mode
Limitations
- Trails Llama 3.1 8B on general chat
- Very enterprise-flavored tone
- Weaker than Qwen 2.5 7B on coding tasks
Typical workloads
In our catalog grid, Granite 3.2 8B Instruct is filed under Enterprise, Lightweight Agents — the use cases where its size/quality trade-off makes the most sense. Its tags translate to concrete workloads: everyday chat, drafting and summarization.
The 125k-token context window covers long chats and mid-sized documents, though very large retrieval workloads will need chunking. The Apache 2.0 license is permissive, so shipping it inside a commercial product raises no special legal questions.
Architecture & training
Architecture: Dense · 8B · IBM Granite 3.2 · RAG and enterprise agents
Training: IBM enterprise corpus, strong in code (100 languages), 2024 data.
The default open 8B for enterprise RAG and regulated workloads — picked for safety guardrails and IBM support, not chat quality.
Quick start
ollama run granite3.2:8bOr use the open-source MCP server to query this model from Claude Desktop, Cursor, or any MCP-compatible client.
Similar models worth comparing
Frequently asked questions
How much VRAM does Granite 3.2 8B Instruct need?
At the recommended Q4_K_M quantization, Granite 3.2 8B Instruct needs about 5 GB of VRAM. Q8_0 takes 9 GB, and unquantized FP16 weights take 16 GB.
Can Granite 3.2 8B Instruct run without a GPU?
Yes — with roughly 10 GB of system RAM it runs CPU-only through llama.cpp or Ollama. Expect a fraction of GPU speed, which is fine for background or batch jobs but slow for interactive chat.
What context window does Granite 3.2 8B Instruct support?
Granite 3.2 8B Instruct supports a 125k-token context window (128,000 tokens).
Can I use Granite 3.2 8B Instruct commercially?
Yes. Granite 3.2 8B Instruct is released under Apache 2.0, a permissive open-source license that allows commercial use, modification and redistribution.
How fast is Granite 3.2 8B Instruct on consumer hardware?
Our compatibility engine estimates on the order of 30 tokens/sec on a mid-range GPU and up to 80 tokens/sec on high-end cards, assuming the quantization fully fits in VRAM.
Which quantization of Granite 3.2 8B Instruct should I download first?
Start with Q4_K_M (5 GB) — the standard size/quality sweet spot. Step up to Q5_K_M or Q8_0 only if you have VRAM headroom. It already fits an 8 GB card at Q5_K_M.