OpenAIGPT2LMHeadModel

Gpt2 Medium — Hardware Requirements & GPU Compatibility

Chat

GPT-2 Medium scales the original GPT-2 architecture to 380 million parameters, offering noticeably improved text generation quality over the base 137M variant while remaining extremely lightweight by current standards. It supports the same autoregressive language modeling tasks as its smaller and larger siblings. Like all GPT-2 variants, it runs comfortably on virtually any modern hardware including CPU-only setups, making it an accessible option for learning, prototyping, and lightweight text generation experiments without needing a dedicated GPU.

380M params · dense1K contextMITMar 2022159.6K downloads208 likes6.1K quant downloads

Your hardware

Verdicts below are for the reference RTX 4090. Set yours to personalise the page.

Detection reads your browser's GPU name with no data sent anywhere.

Runs great on a 24 GB GPU

Runs from 8 GB (Q4_K_M)Best on 24 GB (Q4_K_M · ~2621 tok/s)Macs from 8 GB (Q4_K_M · ~287 tok/s)

MIT

Reference GPU · RTX 4090 24 GB

Estimates, ±20% · how we estimate

Run Gpt2 Medium

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.

Ways to runOther ways

each opens step-by-step

Run in cloud

Runs on RTX 3080 Ti (12 GB) (11 GB headroom) · Q4_K_M

Generation speed: ~2372 tok/s; generation speed
Cost per 1M output tokens: $0.014; per 1M output tokens

Deployon Vast.aiBest valuefrom $0.12/hr

Compare GPUs →

Run Gpt2 Medium

0.3 GB VRAM at 1K context · fits the RTX 4090 (23.7 GB free)

Works on macOS, Windows and Linux.

Install Ollama (Linux)
$curl -fsSL https://ollama.com/install.sh | sh
Install Ollama (macOS / Windows)
Download and run the app; it installs the ollama command.
Download Ollama ↗
Run Q4_K_M
$ollama run hf.co/mradermacher/gpt2-medium-i1-GGUF:Q4_K_M
Raise the context window
Type it inside the chat. For the API server instead, start it with OLLAMA_CONTEXT_LENGTH=1024 ollama serve.
$/set parameter num_ctx 1024
A larger context needs more memory than the weights alone.
Use the API
Ollama serves its API on http://localhost:11434.

Works on macOS, Windows and Linux.

Install LM Studio
Desktop app for macOS (Apple silicon), Windows and Linux.
lmstudio.ai ↗
Download the model
In the app press Ctrl/Cmd + Shift + M and paste https://huggingface.co/mradermacher/gpt2-medium-i1-GGUF, then pick Q4_K_M. Or use the CLI (run LM Studio once first):
$lms get https://huggingface.co/mradermacher/gpt2-medium-i1-GGUF@Q4_K_M
Open on Hugging Face ↗
Load and chat
Open the model loader, select the model and set Context Length to 1024. Leave GPU offload at maximum if it fits.
Start the local server (optional)
$lms server start

Works on macOS, Windows and Linux.

Install llama.cpp (macOS)
MacPorts, Nix and conda-forge also work.
$brew install llama.cpp
Install options ↗
Install llama.cpp (Windows)
$winget install llama.cpp
Install llama.cpp (Linux)
Download the prebuilt llama-*-bin-ubuntu-* archive for your GPU backend (CPU, Vulkan) from the releases page, unpack it and run the binaries from that folder. Homebrew (brew install llama.cpp) and conda-forge also work on Linux.
llama.cpp releases ↗
Start the server
Downloads the model on first run. The chat UI and an OpenAI-compatible API are at http://localhost:8080.
$llama-server -hf mradermacher/gpt2-medium-i1-GGUF:Q4_K_M -c 1024 -ngl 99 --port 8080

Works on macOS, Windows and Linux.

Option A: Docker Model Runner
Docker Desktop: Settings > AI > Enable Docker Model Runner. Docker Engine on Linux: install the docker-model-plugin package.
$docker model run hf.co/mradermacher/gpt2-medium-i1-GGUF:Q4_K_M
Model Runner docs ↗
The quant tag after the repo is how Hugging Face cards write it; if your Docker version rejects it, drop the tag and pick the quant by repo.
Option B: llama.cpp server image (NVIDIA)
Needs the NVIDIA Container Toolkit. Models are cached in the llama-cache volume. Open http://localhost:8080.
$docker run --gpus all -p 8080:8080 -v llama-cache:/root/.cache ghcr.io/ggml-org/llama.cpp:server-cuda -hf mradermacher/gpt2-medium-i1-GGUF:Q4_K_M -c 1024 -ngl 99 --host 0.0.0.0 --port 8080
Image tags (cuda13, vulkan, rocm) ↗
Cache path inside the image is not in the docs; remove the -v flag if the volume complains.

Linux only.

Install vLLM
Linux only; on Windows run it inside WSL2. NVIDIA GPU shown; AMD ROCm, Intel and CPU builds have their own install pages.
$uv venv --python 3.12 --seed
$source .venv/bin/activate
$uv pip install vllm --torch-backend=auto
vLLM installation ↗
Serve the model
OpenAI-compatible API at http://localhost:8000/v1.
$vllm serve openai-community/gpt2-medium --max-model-len 1024
Runs the original weights, not the Q4_K_M GGUF; needs ~1 GB of GPU memory.

macOS only.

MLX supports only macOS, so follow these steps for that setup.

Install mlx-lm
$pip install mlx-lm
Chat in the terminal
$mlx_lm.chat --model viktor2698/gpt2-medium-mlx-8Bit
Downloads the model on first run. The MLX repo has its own quantisation; it may not match the quant selected above.
Or start a local server
Listens on http://localhost:8080.
$mlx_lm.server --model viktor2698/gpt2-medium-mlx-8Bit

Works on macOS, Windows and Linux.

Install Jan
jan.ai ↗
Find the model
Open the Hub tab, paste this model ID in Search, and choose Q4_K_M if offered.
$mradermacher/gpt2-medium-i1-GGUF
Use the local API (optional)
Settings > Local API Server > Start Server. It listens on http://127.0.0.1:1337.

Rent GPUs hourly to run Q4_K_M right there. Pricing comes from partner marketplaces and varies.

Available GPU rentals for Q4_K_M, prioritized by value
GPU	Speed	$/1M tokens	Rent
RTX 3080 Ti (12 GB)CheapestCheapest11 GB free	~2372 tok/s	$0.014	Deploy on Vast.aifrom $0.12/hr (opens in new tab)	12 GB · 11 GB free · ~2372 tok/s · $0.014 / 1M tokens
RTX 3060 12GB11 GB free	~936 tok/s	$0.018	Deploy on Vast.aifrom $0.061/hr (opens in new tab)	11 GB free · ~936 tok/s · $0.018 / 1M tokens
RTX 4060 (8 GB)7 GB free	~707 tok/s	$0.034	Deploy on Vast.aifrom $0.086/hr (opens in new tab)	8 GB · 7 GB free · ~707 tok/s · $0.034 / 1M tokens
RTX A4000 (16 GB)15 GB free	~1165 tok/s	$0.021	Deploy on Vast.aifrom $0.088/hr (opens in new tab)	16 GB · 15 GB free · ~1165 tok/s · $0.021 / 1M tokens
RTX 4070 SUPER (12 GB)11 GB free	~1310 tok/s	$0.022	Deploy on Vast.aifrom $0.10/hr (opens in new tab)	12 GB · 11 GB free · ~1310 tok/s · $0.022 / 1M tokens
RTX PRO 4500 (32 GB)31 GB free	~2330 tok/s	$0.041	Deploy on RunPodfrom $0.34/hr (opens in new tab)	32 GB · 31 GB free · ~2330 tok/s · $0.041 / 1M tokens
RTX 4090 (24 GB)23 GB free	~2621 tok/s	$0.036	Deploy on RunPodfrom $0.34/hr (opens in new tab)	24 GB · 23 GB free · ~2621 tok/s · $0.036 / 1M tokens
A40 (48 GB)47 GB free	~1810 tok/s	$0.054	Deploy on RunPodfrom $0.35/hr (opens in new tab)	48 GB · 47 GB free · ~1810 tok/s · $0.054 / 1M tokens
L4 (24 GB)23 GB free	~780 tok/s	$0.16	Deploy on RunPodfrom $0.44/hr (opens in new tab)	24 GB · 23 GB free · ~780 tok/s · $0.16 / 1M tokens

Compare GPUs →

How Much VRAM Does Gpt2 Medium Need?

Select a quantization level; the listed hardware adheres to it.

VRAM	+32K	+128K	File	Quality	On RTX 4090 (reference)
0.1 GB	—	—	0.1 GB	2-bit · i-quant extreme compression — Importance-weighted 2-bit, extreme compression — significant quality loss	Fits: — context · ~5956 tok/s
0.1 GB	—	—	0.1 GB	2-bit · i-quant extra small — Importance-weighted 2-bit, extra small	Fits: — context · ~5040 tok/s
0.1 GB	—	—	0.2 GB	2-bit · i-quant small — Importance-weighted 2-bit, small	Fits: — context · ~5040 tok/s
0.1 GB	—	—	0.2 GB	2-bit · i-quant medium — Importance-weighted 2-bit, medium	Fits: — context · ~4680 tok/s
0.2 GB	—	—	0.2 GB	3-bit · i-quant — Importance-weighted 3-bit	Fits: — context · ~4095 tok/s
0.2 GB	—	—	0.2 GB	3-bit · i-quant extra small — Importance-weighted 3-bit, extra small	Fits: — context · ~3854 tok/s
0.2 GB	—	—	0.2 GB	2-bit · K-quant — 2-bit quantization with K-quant improvements	Fits: — context · ~3640 tok/s
0.2 GB	—	—	0.2 GB	3-bit small — 3-bit small quantization	Fits: — context · ~3640 tok/s
0.2 GB	—	—	0.2 GB	3-bit · i-quant small — Importance-weighted 3-bit, small	Fits: — context · ~3640 tok/s
0.2 GB	—	—	0.2 GB	3-bit · i-quant medium — Importance-weighted 3-bit, medium	Fits: — context · ~3448 tok/s
0.2 GB	—	—	0.2 GB	3-bit medium — 3-bit medium quantization	Fits: — context · ~3276 tok/s
0.2 GB	—	—	0.2 GB	4-bit · legacy — 4-bit legacy quantization	Fits: — context · ~3120 tok/s
0.2 GB	—	—	0.2 GB	3-bit large — 3-bit large quantization	Fits: — context · ~3120 tok/s
0.2 GB	—	—	0.2 GB	4-bit · i-quant compact — Importance-weighted 4-bit, compact	Fits: — context · ~2978 tok/s
0.2 GB	—	—	0.2 GB	4-bit small — 4-bit small quantization	Fits: — context · ~2730 tok/s
0.3 GB	—	—	0.3 GB	4-bit medium — 4-bit medium quantization — most popular sweet spot	Fits: — context · ~2621 tok/s
0.3 GB	—	—	0.3 GB	5-bit small — 5-bit small quantization	Fits: — context · ~2259 tok/s
0.3 GB	—	—	0.3 GB	5-bit medium — 5-bit medium quantization — good quality/size tradeoff	Fits: — context · ~2184 tok/s
0.3 GB	—	—	0.3 GB	6-bit, very good — 6-bit quantization, very good quality	Fits: — context · ~1927 tok/s
0.4 GB	—	—	0.4 GB	8-bit, near-lossless — 8-bit quantization, near-lossless	Fits: — context · ~1560 tok/s
0.8 GB	—	—	0.8 GB	16-bit · bfloat — Brain floating point 16 — preferred for training	Fits: — context · ~780 tok/s

Selected Q4_K_M

4-bit medium quantization

0.3GBat 1K context

Weights: 0.1 GB
KV cache: 0.0 GB
Overhead: 0.3 GB

Fits RTX 4090 (reference) · 23.8 GB free · ~2621 tok/s

≥50 tok/s needs ≥19 GB/s · RTX 5090 ~4659, RTX 3090 Ti ~2621

Which Hardware Can Run Gpt2 Medium?

Q4_K_M · 0.3 GB

Gpt2 Medium (Q4_K_M) requires 0.3 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 1+ GB is recommended. 52 GPUs can run it, including NVIDIA GeForce RTX 5090, NVIDIA GeForce RTX 3090 Ti. 163 devices with unified memory can run Gpt2 Medium, including Mac Studio (M3 Ultra, 256GB), Mac Studio (M3 Ultra, 512GB).

By VRAM tier · typical GPUs and Macs at Q4_K_M

Typical hardwareQuantVRAMtok/sContext that fits

8 GB · RTX 4060Q4_K_M · 0.3 GB · ~707 tok/s · —0.3~707—8 GB tier →

12 GB · RTX 3060 12GBQ4_K_M · 0.3 GB · ~936 tok/s · —0.3~936—12 GB tier →

16 GB · RTX 5060 Ti 16GBQ4_K_M · 0.3 GB · ~1165 tok/s · —0.3~1165—16 GB tier →

24 GB · RTX 4090referenceQ4_K_M · 0.3 GB · ~2621 tok/s · —0.3~2621—24 GB tier →

Mac 32 GB · Mac Mini M4Q4_K_M · 0.3 GB · ~336 tok/s · —0.3~336—Macs →

Mac 64 GB · Mac Studio M4 MaxQ4_K_M · 0.3 GB · ~1529 tok/s · —0.3~1529—Macs →

Alternatives to Gpt2 Medium

Similar models

5tok/s on RTX 4090

Model	Score	VRAM	tok/s	Why it is similar
Granite 4.0 350MIBM	—	1 GB	~1150	same size
SmolLM 360MHugging Face	—	0.6 GB	~1092	same size
LFM2.5 350MLiquid AI	—	0.8 GB	~1130	same size
ERNIE 4.5 0.3B PTBaidu	—	0.7 GB	~1213	same size
LFM2 350MLiquid AI	—	0.8 GB	~1130	same size

Fine-tunes & variants(1)community builds of Gpt2 Medium

MiniLLM Gpt2 340M

Specifications

Publisher: OpenAI
Parameters: 380M
Architecture: GPT2LMHeadModel
Context Length: 1,024 tokens
Vocabulary Size: 50,257
Release Date: 2022-03-02
License: MIT

Resources

Hugging Face model card (opens in a new tab)

Frequently Asked Questions

How much VRAM does Gpt2 Medium need?

Gpt2 Medium requires 0.3 GB of VRAM at Q4_K_M, or 0.8 GB at BF16.

VRAM = Weights + KV Cache + Overhead

Weights = 380M × 4.8 bits ÷ 8 = 0.2 GB

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

Q4_K_M

0.3 GB

Learn more about VRAM estimation →

What's the best quantization for Gpt2 Medium?

For Gpt2 Medium, Q4_K_M (0.3 GB) offers the best balance of quality and VRAM usage. Q5_K_S (0.3 GB) provides better quality if you have the VRAM. The smallest option is IQ2_XXS at 0.1 GB.

VRAM requirement by quantization

IQ2_XXS

0.1 GB~53%

IQ3_XS

0.2 GB~73%

Q3_K_M

0.2 GB~83%

Q4_K_M ★

0.3 GB~89%

Q5_K_S

0.3 GB~92%

BF16

0.8 GB~100%

★ Recommended — best balance of quality and VRAM usage.

Learn more about quantization →

Can I run Gpt2 Medium on a Mac?

Yes — MacBook Air 13" M3 (8 GB) and 145 other Macs can run Gpt2 Medium. Apple Silicon uses unified memory, so the model shares RAM with the system. At Q4_K_M you need at least 0.3 GB of usable unified memory (RAM minus macOS overhead).

Can I run Gpt2 Medium locally?

Yes — Gpt2 Medium can run locally on consumer hardware. Popular tools include Ollama, LM Studio, and llama.cpp.

How fast is Gpt2 Medium?

At Q4_K_M, Gpt2 Medium can reach ~19200 tok/s on AMD Instinct MI350X. On NVIDIA GeForce RTX 4090: ~2621 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 ÷ 0.25 × 0.65 = ~20800 tok/s

Estimated speed at Q4_K_M (0.3 GB)

NVIDIA B200

~20800 tok/s

NVIDIA GeForce RTX 4090

~2621 tok/s

NVIDIA B300

~20800 tok/s

AMD Instinct MI350X

~19200 tok/s

Real-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.

Learn more about tok/s estimation →

What's the download size of Gpt2 Medium?

At Q4_K_M, the download is about 0.23 GB. The full-precision BF16 version is 0.76 GB. The smallest option (IQ2_XXS) is 0.10 GB.

Which GPUs can run Gpt2 Medium?

52 consumer GPUs can run Gpt2 Medium at Q4_K_M. Top options include AMD Radeon RX 6700 XT, AMD Radeon RX 6800, AMD Radeon RX 6800 XT. 52 GPUs have plenty of headroom for comfortable inference.

Which devices can run Gpt2 Medium?

163 devices with unified memory can run Gpt2 Medium at Q4_K_M (0.3 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.