Opt 6.7B — Hardware Requirements & GPU Compatibility
Chat
Opt 6.7B is a 6.7B-parameter open language model from Meta. It supports a context window of up to 2,048 tokens. At FP16 it needs about 14.74 GB of VRAM — see which GPUs and Macs can run it below.
Runs well on a 24 GB GPU
Runs from 16 GB (FP16 · 2K)Best on 24 GB (FP16 · ~45 tok/s · 2K)Macs from 24 GB (FP16 · ~15 tok/s)
Other
Reference GPU · RTX 4090 24 GB
Estimates, ±20% · how we estimate
Run Opt 6.7B
One recommended way. Every other method, OS and quant is in the step-by-step guide.Each way opens a step-by-step guide for your OS and quant.
Recommended
vllm serve facebook/opt-6.7b --max-model-len 2048vLLM repo with its own quantisation
Download: Weights on Hugging Face
Ways to runOther ways
each opens step-by-stepRun in cloud
Runs on RTX 3090 (24 GB) (9 GB headroom) · FP16
- Generation speed
- ~41 tok/s
- generation speed
- Cost per 1M output tokens
- $1.33
- per 1M output tokens
How Much VRAM Does Opt 6.7B Need?
Select a quantization level; the listed hardware adheres to it.
| Quantization | VRAM | +32K | On RTX 4090 (reference) | |
|---|---|---|---|---|
| 14.7 GB | — | 16-bit · full — Full half-precision — baseline for inference | Fits: 2K context · ~45 tok/s |
Selected FP16
Full half-precision14.7GBat 2K context
- Weights
- 14.1 GB
- KV cache
- 0.4 GB
- Overhead
- 0.3 GB
Fits RTX 4090 (reference) · 9.3 GB free · ~45 tok/s
≥50 tok/s needs ≥1.13 TB/s · RTX 5090 ~79
Which Hardware Can Run Opt 6.7B?
FP16 · 14.7 GBOpt 6.7B (FP16) requires 14.7 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 20+ GB is recommended. 26 GPUs can run it, including NVIDIA GeForce RTX 5090, NVIDIA GeForce RTX 3090 Ti, NVIDIA GeForce RTX 5080. 122 devices with unified memory can run Opt 6.7B, including Mac Studio (M3 Ultra, 256GB), Mac Studio (M3 Ultra, 512GB), Mac Mini M5 Pro (24 GB).
16–18 GB
16–18 GB tier →fits · 1.3 GB free · up to 2K ctx
18 GPUs
RTX 5080~42Simulate output on RTX 5080RTX 5070 Ti~40Simulate output on RTX 5070 TiRTX 4080 SUPER~33Simulate output on RTX 4080 SUPER20 GB
fits · 5.3 GB free · up to 2K ctx
1 GPU
RX 7900 XT~33Simulate output on RX 7900 XT32–40 GB
fits · 17.3 GB free · up to 2K ctx
1 GPU · 3 Mac classes · 4 device families
RTX 5090~79Simulate output on RTX 5090M5 Max36 GB~22Simulate output on Mac Studio M5 Max (36 GB)M532 GB~7Simulate output on MacBook Air 13" M5 (32 GB)Orin~9Simulate output on NVIDIA Jetson AGX Orin 32GBSnapdragon X Elite (X1E-80-100)~6Simulate output on Snapdragon X Elite Copilot+ PC
By VRAM tier · typical GPUs and Macs at FP16
Similar models
6| Model | Score | VRAM | tok/s | Why it is similar |
|---|---|---|---|---|
| Llava 1.5 7B HFllava-hf | — | 5 GB | ~141 | same size |
| Chatglm3 6BZ.ai | — | 5.2 GB | ~159 | same size |
| Baichuan2 7B Basebaichuan-inc | — | 5 GB | ~142 | same size |
| Internlm 7BInternLM | — | 15.4 GB | ~43 | same size |
| Olmo Hybrid 7BAllen AI | — | 22.4 GB | ~43 | same size |
| Granite 4.0 H TinyIBM | — | 5.8 GB | ~271 | same size |
Specifications
- Publisher
- Meta
- Parameters
- 6.7B
- Architecture
- OPTForCausalLM
- Context Length
- 2,048 tokens
- Vocabulary Size
- 50,272
- Release Date
- 2022-05-11
- License
- Other
Frequently Asked Questions
How much VRAM does Opt 6.7B need?
Opt 6.7B requires 14.7 GB of VRAM at FP16.
VRAM = Weights + KV Cache + Overhead
Weights = 6.7B × 16 bits ÷ 8 = 13.4 GB
KV Cache + Overhead ≈ 1.3 GB (at 2K context + ~0.3 GB framework)
Fit ratings and hardware model lists check this model with room for a 16K-token context, which needs a little more memory.
VRAM usage by quantization
Can I run Opt 6.7B on a Mac?
Yes — Mac Mini M5 Pro (24 GB) and 106 other Macs can run Opt 6.7B. Apple Silicon uses unified memory, so the model shares RAM with the system. At FP16 you need at least 14.7 GB of usable unified memory (RAM minus macOS overhead).
Can I run Opt 6.7B locally?
Yes — Opt 6.7B can run locally on consumer hardware. Popular tools include Ollama, LM Studio, and llama.cpp.
How fast is Opt 6.7B?
At FP16, Opt 6.7B can reach ~326 tok/s on AMD Instinct MI350X. On NVIDIA GeForce RTX 4090: ~45 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 ÷ 14.74 × 0.65 = ~353 tok/s
Estimated speed at FP16 (14.7 GB)
Real-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
What's the download size of Opt 6.7B?
At FP16, the download is about 13.40 GB.
Which GPUs can run Opt 6.7B?
26 consumer GPUs can run Opt 6.7B at FP16. 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 Opt 6.7B?
122 devices with unified memory can run Opt 6.7B at FP16 (14.7 GB), including AMD Ryzen AI 9 HX 370 (Strix Point) Laptop, ASUS Ascent GX10, Beelink GTR9 Pro (Ryzen AI Max+ 395, 128 GB), Framework Desktop (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.