Clef — Hardware Requirements & GPU Compatibility
Vision
Clef is Cloudflare's 27-billion-parameter multimodal decision model, post-trained from Qwen3.8-27B. Unlike a chat model, it does not generate free-form text. It reads a state given as text, JSON, images or video, plus a schema of typed questions, and returns a probability for every allowed option in a single forward pass through a small joint schema head. It suits classification, routing and structured decisions. The card says it was tested on a single H200, and at this size it fits on one high-memory GPU once quantized, though the custom head needs the repository's own code. The context length is 262,144 tokens. It is released under the Apache 2.0 license, permitting unrestricted commercial and research use. Published in September 2026, it has a smaller sibling, Clef-Flash, built on a 9B Qwen3.5 backbone.
824 downloads
803 likesSpecifications
- Publisher
- Cloudflare
- Parameters
- 27.4B
- Architecture
- Qwen3_5ForConditionalGeneration
- Context Length
- 262,144 tokens
- Vocabulary Size
- 248,320
- Release Date
- 2026-09-30
- License
- Apache 2.0
Get Started
HuggingFace
How Much VRAM Does Clef Need?
Select a quantization to see compatible GPUs below.
| Quantization | Bits | VRAM | + Context | File Size | Quality |
|---|---|---|---|---|---|
| Q2_K | 3.40 | 12.4 GB | 69.2 GB | 11.63 GB | 2-bit quantization with K-quant improvements |
| Q3_K_S | 3.50 | 12.7 GB | 69.5 GB | 11.97 GB | 3-bit small quantization |
| Q3_K_M | 3.90 | 14.1 GB | 70.9 GB | 13.34 GB | 3-bit medium quantization |
| Q4_0 | 4.00 | 14.4 GB | 71.2 GB | 13.68 GB | 4-bit legacy quantization |
| Q4_K_M | 4.80 | 17.2 GB | 74.0 GB | 16.41 GB | 4-bit medium quantization — most popular sweet spot |
| Q5_K_M | 5.70 | 20.2 GB | 77.1 GB | 19.49 GB | 5-bit medium quantization — good quality/size tradeoff |
| Q6_K | 6.60 | 23.3 GB | 80.1 GB | 22.57 GB | 6-bit quantization, very good quality |
| Q8_0 | 8.00 | 28.1 GB | 84.9 GB | 27.36 GB | 8-bit quantization, near-lossless |
Which GPUs Can Run Clef?
Q4_K_M · 17.2 GBClef (Q4_K_M) requires 17.2 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 23+ GB is recommended. Using the full 262K context window can add up to 56.8 GB, bringing total usage to 74.0 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 Clef?
Q4_K_M · 17.2 GB41 devices with unified memory can run Clef, 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 Clef
Community quantizations of this model — GGUF for llama.cpp, Ollama, and LM Studio, plus AWQ/MLX variants where available.
Related Models
Frequently Asked Questions
- How much VRAM does Clef need?
Clef requires 17.2 GB of VRAM at Q4_K_M, or 55.5 GB at BF16. Full 262K context adds up to 56.8 GB (74.0 GB total).
VRAM = Weights + KV Cache + Overhead
Weights = 27.4B × 4.8 bits ÷ 8 = 16.4 GB
KV Cache + Overhead ≈ 0.8 GB (at 2K context + ~0.3 GB framework)
KV Cache + Overhead ≈ 57.6 GB (at full 262K context)
VRAM usage by quantization
Q4_K_M17.2 GBQ4_K_M + full context74.0 GB- Can NVIDIA GeForce RTX 4090 run Clef?
Yes, at Q6_K (23.3 GB) or lower. Higher quantizations like Q8_0 (28.1 GB) exceed the NVIDIA GeForce RTX 4090's 24 GB.
- What's the best quantization for Clef?
For Clef, Q4_K_M (17.2 GB) offers the best balance of quality and VRAM usage. Q4_K_L (17.5 GB) provides better quality if you have the VRAM. The smallest option is IQ2_XXS at 8.3 GB.
VRAM requirement by quantization
IQ2_XXS8.3 GBQ2_K12.4 GBQ3_K_L14.8 GBQ4_K_M ★17.2 GBQ4_K_L17.5 GBBF1655.5 GB★ Recommended — best balance of quality and VRAM usage.
- Can I run Clef on a Mac?
Clef requires at least 8.3 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 Clef locally?
Yes — Clef can run locally on consumer hardware. At Q4_K_M quantization it needs 17.2 GB of VRAM. Popular tools include Ollama, LM Studio, and llama.cpp.
- How fast is Clef?
At Q4_K_M, Clef can reach ~280 tok/s on AMD Instinct MI350X. On NVIDIA GeForce RTX 4090: ~38 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 ÷ 17.2 × 0.65 = ~303 tok/s
Estimated speed at Q4_K_M (17.2 GB)
~303 tok/s~38 tok/s~303 tok/s~280 tok/sReal-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
- What's the download size of Clef?
At Q4_K_M, the download is about 16.41 GB. The full-precision BF16 version is 54.71 GB. The smallest option (IQ2_XXS) is 7.52 GB.
- Which GPUs can run Clef?
8 consumer GPUs can run Clef at Q4_K_M (17.2 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 Clef?
41 devices with unified memory can run Clef at Q4_K_M (17.2 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.