EuroLLM 22B Instruct 2512 — Hardware Requirements & GPU Compatibility
Chat
EuroLLM 22B Instruct 2512 is a 22.6B-parameter open language model from utter-project. It supports a context window of up to 32,768 tokens. At Q4_K_M it needs about 14.34 GB of VRAM — see which GPUs and Macs can run it below.
296.6K downloads
83 likesSpecifications
- Publisher
- utter-project
- Parameters
- 22.6B
- Architecture
- LlamaForCausalLM
- Context Length
- 32,768 tokens
- Vocabulary Size
- 128,000
- Release Date
- 2025-12-05
- License
- Apache 2.0
Get Started
HuggingFace
How Much VRAM Does EuroLLM 22B Instruct 2512 Need?
Select a quantization to see compatible GPUs below.
| Quantization | Bits | VRAM | + Context | File Size | Quality |
|---|---|---|---|---|---|
| Q2_K | 3.40 | 10.4 GB | 17.2 GB | 9.62 GB | 2-bit quantization with K-quant improvements |
| Q3_K_S | 3.50 | 10.7 GB | 17.4 GB | 9.90 GB | 3-bit small quantization |
| Q3_K_M | 3.90 | 11.8 GB | 18.6 GB | 11.04 GB | 3-bit medium quantization |
| Q4_0 | 4.00 | 12.1 GB | 18.9 GB | 11.32 GB | 4-bit legacy quantization |
| Q4_K_M | 4.80 | 14.3 GB | 21.1 GB | 13.58 GB | 4-bit medium quantization — most popular sweet spot |
| Q5_K_M | 5.70 | 16.9 GB | 23.7 GB | 16.13 GB | 5-bit medium quantization — good quality/size tradeoff |
| Q6_K | 6.60 | 19.4 GB | 26.2 GB | 18.68 GB | 6-bit quantization, very good quality |
| Q8_0 | 8.00 | 23.4 GB | 30.2 GB | 22.64 GB | 8-bit quantization, near-lossless |
Which GPUs Can Run EuroLLM 22B Instruct 2512?
Q4_K_M · 14.3 GBEuroLLM 22B Instruct 2512 (Q4_K_M) requires 14.3 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 19+ GB is recommended. Using the full 33K context window can add up to 6.8 GB, bringing total usage to 21.1 GB. 26 GPUs can run it, including NVIDIA GeForce RTX 5090, NVIDIA GeForce RTX 3090 Ti, NVIDIA GeForce RTX 5080.
Runs great
— Plenty of headroom
Decent
— Enough VRAM, may be tightWhich Devices Can Run EuroLLM 22B Instruct 2512?
Q4_K_M · 14.3 GB47 devices with unified memory can run EuroLLM 22B Instruct 2512, including NVIDIA DGX H100, NVIDIA DGX A100 640GB, Mac Mini M4 (16 GB).
Runs great
— Plenty of headroomWhere to Download EuroLLM 22B Instruct 2512
Community quantizations of this model — GGUF for llama.cpp, Ollama, and LM Studio, plus AWQ/MLX variants where available.
Frequently Asked Questions
- How much VRAM does EuroLLM 22B Instruct 2512 need?
EuroLLM 22B Instruct 2512 requires 14.3 GB of VRAM at Q4_K_M, or 46.0 GB at BF16. Full 33K context adds up to 6.8 GB (21.1 GB total).
VRAM = Weights + KV Cache + Overhead
Weights = 22.6B × 4.8 bits ÷ 8 = 13.6 GB
KV Cache + Overhead ≈ 0.7 GB (at 2K context + ~0.3 GB framework)
KV Cache + Overhead ≈ 7.5 GB (at full 33K context)
VRAM usage by quantization
Q4_K_M14.3 GBQ4_K_M + full context21.1 GB- Can NVIDIA GeForce RTX 4090 run EuroLLM 22B Instruct 2512?
Yes, at Q8_0 (23.4 GB) or lower. Higher quantizations like BF16 (46.0 GB) exceed the NVIDIA GeForce RTX 4090's 24 GB.
- What's the best quantization for EuroLLM 22B Instruct 2512?
For EuroLLM 22B Instruct 2512, Q4_K_M (14.3 GB) offers the best balance of quality and VRAM usage. Q4_K_L (14.6 GB) provides better quality if you have the VRAM. The smallest option is IQ2_XXS at 7.0 GB.
VRAM requirement by quantization
IQ2_XXS7.0 GBIQ3_XS10.1 GBQ3_K_L12.3 GBQ4_K_M ★14.3 GBQ4_K_L14.6 GBBF1646.0 GB★ Recommended — best balance of quality and VRAM usage.
- Can I run EuroLLM 22B Instruct 2512 on a Mac?
EuroLLM 22B Instruct 2512 requires at least 7.0 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 EuroLLM 22B Instruct 2512 locally?
Yes — EuroLLM 22B Instruct 2512 can run locally on consumer hardware. At Q4_K_M quantization it needs 14.3 GB of VRAM. Popular tools include Ollama, LM Studio, and llama.cpp.
- How fast is EuroLLM 22B Instruct 2512?
At Q4_K_M, EuroLLM 22B Instruct 2512 can reach ~335 tok/s on AMD Instinct MI350X. On NVIDIA GeForce RTX 4090: ~46 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 ÷ 14.3 × 0.65 = ~363 tok/s
Estimated speed at Q4_K_M (14.3 GB)
~363 tok/s~46 tok/s~363 tok/s~335 tok/sReal-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
- What's the download size of EuroLLM 22B Instruct 2512?
At Q4_K_M, the download is about 13.58 GB. The full-precision BF16 version is 45.27 GB. The smallest option (IQ2_XXS) is 6.23 GB.
- Which GPUs can run EuroLLM 22B Instruct 2512?
26 consumer GPUs can run EuroLLM 22B Instruct 2512 at Q4_K_M (14.3 GB). Top options include AMD Radeon RX 7900 XTX, NVIDIA GeForce RTX 3090, NVIDIA GeForce RTX 3090 Ti, AMD Radeon RX 6800. 7 GPUs have plenty of headroom for comfortable inference.
- Which devices can run EuroLLM 22B Instruct 2512?
49 devices with unified memory can run EuroLLM 22B Instruct 2512 at Q4_K_M (14.3 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.