Allen AIOLMoOlmo3ForCausalLM

Olmo 3 1025 7B — Hardware Requirements & GPU Compatibility

Chat

Olmo 3 1025 7B is a 7.3B-parameter open language model from Allen AI in the OLMo family. It supports a context window of up to 65,536 tokens. At Q4_K_M it needs about 5.75 GB of VRAM — see which GPUs and Macs can run it below.

7.3B params · dense66K contextApache 2.0Sep 2025157.2K downloads89 likes1.2K 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 · 8K)Best on 24 GB (Q4_K_M · ~114 tok/s · 64K)Macs from 16 GB (Q4_K_M · ~24 tok/s)

Apache 2.0

Reference GPU · RTX 4090 24 GB

Estimates, ±20% · how we estimate

Run Olmo 3 1025 7B

One recommended way. Every other method, OS and quant is in the step-by-step guide.

Other ways

each opens step-by-step

Run in cloud

Fits on RTX 3060 12GB (3 GB headroom) · Q4_K_M

Generation speed: ~41 tok/s; generation speed
Cost per 1M output tokens: $0.47; per 1M output tokens

Deployon Vast.aiBest optionfrom $0.069/hr

Compare GPUs →

Run Olmo 3 1025 7B

8.4 GB VRAM at 16K context · fits the RTX 4090 (15.6 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/Olmo-3-1025-7B-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=16384 ollama serve.
$/set parameter num_ctx 16384
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/Olmo-3-1025-7B-i1-GGUF, then pick Q4_K_M. Or use the CLI (run LM Studio once first):
$lms get https://huggingface.co/mradermacher/Olmo-3-1025-7B-i1-GGUF@Q4_K_M
Open on Hugging Face ↗
Load and chat
Open the model loader, select the model and set Context Length to 16384. 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/Olmo-3-1025-7B-i1-GGUF:Q4_K_M -c 16384 -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/Olmo-3-1025-7B-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/Olmo-3-1025-7B-i1-GGUF:Q4_K_M -c 16384 -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 allenai/Olmo-3-1025-7B --max-model-len 16384
Runs the original weights, not the Q4_K_M GGUF; needs ~16 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 lmstudio-community/Olmo-3-1025-7B-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 lmstudio-community/Olmo-3-1025-7B-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/Olmo-3-1025-7B-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 3060 12GBCheapestCheapest3 GB free	~41 tok/s	$0.47	Deploy on Vast.aifrom $0.069/hr (opens in new tab)	3 GB free · ~41 tok/s · $0.47 / 1M tokens
Tesla V100 PCIe 16GB7 GB free	~102 tok/s	$0.26	Deploy on Vast.aifrom $0.096/hr (opens in new tab)	7 GB free · ~102 tok/s · $0.26 / 1M tokens
T4 (16 GB)7 GB free	~36 tok/s	$1.15	Deploy on Vast.aifrom $0.15/hr (opens in new tab)	16 GB · 7 GB free · ~36 tok/s · $1.15 / 1M tokens
RTX A5000 (24 GB)15 GB free	~87 tok/s	$0.51	Deploy on RunPodfrom $0.16/hr (opens in new tab)	24 GB · 15 GB free · ~87 tok/s · $0.51 / 1M tokens
RTX 4060 Ti 16GB7 GB free	~33 tok/s	$1.38	Deploy on Vast.aifrom $0.16/hr (opens in new tab)	7 GB free · ~33 tok/s · $1.38 / 1M tokens
RTX 4000 Ada (20 GB)11 GB free	~41 tok/s	$1.37	Deploy on RunPodfrom $0.20/hr (opens in new tab)	20 GB · 11 GB free · ~41 tok/s · $1.37 / 1M tokens
RTX 3090 (24 GB)15 GB free	~106 tok/s	$0.58	Deploy on RunPodfrom $0.22/hr (opens in new tab)	24 GB · 15 GB free · ~106 tok/s · $0.58 / 1M tokens
RTX A6000 (48 GB)39 GB free	~87 tok/s	$1.06	Deploy on RunPodfrom $0.33/hr (opens in new tab)	48 GB · 39 GB free · ~87 tok/s · $1.06 / 1M tokens

Compare GPUs →

How Much VRAM Does Olmo 3 1025 7B Need?

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

VRAM	+32K	+128K	File	Quality	On RTX 4090 (reference)
3.4 GB	8.2 GB	—	2.1 GB	2-bit · i-quant extreme compression — Importance-weighted 2-bit, extreme compression — significant quality loss	Fits: 64K context · ~194 tok/s
3.6 GB	8.4 GB	—	2.3 GB	2-bit · i-quant extra small — Importance-weighted 2-bit, extra small	Fits: 64K context · ~184 tok/s
3.6 GB	8.5 GB	—	2.5 GB	2-bit · i-quant small — Importance-weighted 2-bit, small	Fits: 64K context · ~180 tok/s
3.8 GB	8.7 GB	—	2.7 GB	2-bit · i-quant medium — Importance-weighted 2-bit, medium	Fits: 64K context · ~171 tok/s
4.2 GB	9 GB	—	2.9 GB	3-bit · i-quant — Importance-weighted 3-bit	Fits: 64K context · ~156 tok/s
4.3 GB	9.1 GB	—	2.6 GB	2-bit · small K-quant — 2-bit small K-quant	Fits: 64K context · ~153 tok/s
4.4 GB	9.2 GB	—	3.1 GB	3-bit · i-quant extra small — Importance-weighted 3-bit, extra small	Fits: 64K context · ~150 tok/s
4.5 GB	9.3 GB	—	3.3 GB	3-bit · i-quant small — Importance-weighted 3-bit, small	Fits: 64K context · ~146 tok/s
4.5 GB	9.3 GB	—	2.9 GB	2-bit · K-quant — 2-bit quantization with K-quant improvements	Fits: 64K context · ~146 tok/s
4.6 GB	9.4 GB	—	3.3 GB	3-bit small — 3-bit small quantization	Fits: 64K context · ~143 tok/s
4.7 GB	9.5 GB	—	3.5 GB	3-bit · i-quant medium — Importance-weighted 3-bit, medium	Fits: 64K context · ~141 tok/s
4.9 GB	9.8 GB	—	3.7 GB	3-bit medium — 3-bit medium quantization	Fits: 64K context · ~133 tok/s
5.0 GB	9.9 GB	—	4.2 GB	4-bit · legacy — 4-bit legacy quantization	Fits: 64K context · ~131 tok/s
5.1 GB	9.9 GB	—	4 GB	3-bit large — 3-bit large quantization	Fits: 64K context · ~128 tok/s
5.3 GB	10.1 GB	—	4 GB	4-bit · i-quant compact — Importance-weighted 4-bit, compact	Fits: 64K context · ~124 tok/s
5.5 GB	10.3 GB	—	4.6 GB	4-bit · legacy+ — 4-bit legacy quantization with offset	Fits: 64K context · ~120 tok/s
5.5 GB	10.3 GB	—	4.2 GB	4-bit small — 4-bit small quantization	Fits: 64K context · ~120 tok/s
5.5 GB	10.3 GB	—	4.2 GB	4-bit · i-quant non-linear — Importance-weighted 4-bit, non-linear	Fits: 64K context · ~120 tok/s
5.8 GB	10.6 GB	—	4.5 GB	4-bit medium — 4-bit medium quantization — most popular sweet spot	Fits: 64K context · ~114 tok/s
6.4 GB	11.2 GB	—	5.1 GB	5-bit small — 5-bit small quantization	Fits: 64K context · ~103 tok/s
6.6 GB	11.4 GB	—	5.2 GB	5-bit medium — 5-bit medium quantization — good quality/size tradeoff	Fits: 64K context · ~100 tok/s
7.4 GB	12.2 GB	—	6 GB	6-bit, very good — 6-bit quantization, very good quality	Fits: 64K context · ~89 tok/s
8.7 GB	13.5 GB	—	7.8 GB	8-bit, near-lossless — 8-bit quantization, near-lossless	Fits: 64K context · ~76 tok/s
16.0 GB	20.8 GB	—	14.6 GB	16-bit · bfloat — Brain floating point 16 — preferred for training	Fits: 32K context · ~41 tok/s

Selected Q4_K_M

4-bit medium quantization

8.4GBat 16K context

Weights: 4.4 GB
KV cache: 3.8 GB
Overhead: 0.3 GB

Fits RTX 4090 (reference) · 15.6 GB free · ~114 tok/s

≥50 tok/s needs ≥442 GB/s · RTX 5090 ~203, RTX 3090 Ti ~114

Which Hardware Can Run Olmo 3 1025 7B?

Q4_K_M · 5.8 GB

Olmo 3 1025 7B (Q4_K_M) requires 5.8 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 8+ GB is recommended. Using the full 66K context window can add up to 9.1 GB, bringing total usage to 14.9 GB. 52 GPUs can run it, including NVIDIA GeForce RTX 5090, NVIDIA GeForce RTX 3090 Ti, NVIDIA GeForce RTX 3070 Ti. 151 devices with unified memory can run Olmo 3 1025 7B, including Mac Studio (M3 Ultra, 256GB), Mac Studio (M3 Ultra, 512GB), NVIDIA Jetson Orin Nano 8GB (Super).

By VRAM tier · typical GPUs and Macs at Q4_K_M

Typical hardwareQuantVRAMtok/sContext that fits

8 GB · RTX 4060Q4_K_M · 8.4 GB · ~31 tok/s · 8Ktight8.4~318K8 GB tier →

12 GB · RTX 3060 12GBQ4_K_M · 8.4 GB · ~41 tok/s · 32K8.4~4132K12 GB tier →

16 GB · RTX 5060 Ti 16GBQ4_K_M · 8.4 GB · ~51 tok/s · 64K8.4~5164K16 GB tier →

24 GB · RTX 4090referenceQ4_K_M · 8.4 GB · ~114 tok/s · 64K8.4~11464K24 GB tier →

Mac 32 GB · Mac Mini M4Q4_K_M · 8.4 GB · ~15 tok/s · 64K8.4~1564KMacs →

Mac 64 GB · Mac Studio M4 MaxQ4_K_M · 8.4 GB · ~67 tok/s · 64K8.4~6764KMacs →

Alternatives to Olmo 3 1025 7B

OLMo family · 20 models →

Similar models

6tok/s on RTX 4090

Model	Score	VRAM	tok/s	Why it is similar
Qwen2.5 7B InstructAlibaba	65.1	5.8 GB	~131	same size
Phi Mini MoE InstructMicrosoft	—	5.4 GB	~239	same size
Gemma 3n E4B ITGoogle	—	7.7 GB	~127	same size
Falcon H1 7B InstructTII UAE	—	6.3 GB	~130	same size
DeepSeek R1 Distill Qwen 7BDeepSeek	—	5.8 GB	~131	same size
Eurus 2 7B PRIMEPRIME-RL	—	5.1 GB	~131	same size

Fine-tunes & variants(2)community builds of Olmo 3 1025 7B

Specifications

Publisher: Allen AI
Family: OLMo
Parameters: 7.3B
Architecture: Olmo3ForCausalLM
Context Length: 65,536 tokens
Vocabulary Size: 100,278
Release Date: 2025-09-12
License: Apache 2.0

Resources

Frequently Asked Questions

How much VRAM does Olmo 3 1025 7B need?

Olmo 3 1025 7B requires 5.8 GB of VRAM at Q4_K_M, or 16.0 GB at BF16. Full 66K context adds up to 9.1 GB (14.9 GB total).

VRAM = Weights + KV Cache + Overhead

Weights = 7.3B × 4.8 bits ÷ 8 = 4.4 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.

KV Cache + Overhead ≈ 10.5 GB (at full 66K context)

VRAM usage by quantization

Q4_K_M

5.8 GB

Q4_K_M + full context

14.9 GB

Learn more about VRAM estimation →

What's the best quantization for Olmo 3 1025 7B?

For Olmo 3 1025 7B, Q4_K_M (5.8 GB) offers the best balance of quality and VRAM usage. Q5_K_S (6.4 GB) provides better quality if you have the VRAM. The smallest option is IQ2_XXS at 3.4 GB.

VRAM requirement by quantization

IQ2_XXS

3.4 GB~53%

IQ3_XS

4.4 GB~73%

Q4_0

5.0 GB~85%

IQ4_NL

5.5 GB~88%

Q4_K_M ★

5.8 GB~89%

BF16

16.0 GB~100%

★ Recommended — best balance of quality and VRAM usage.

Learn more about quantization →

Can I run Olmo 3 1025 7B on a Mac?

Yes — Mac Mini M2 Pro (16 GB) and 133 other Macs can run Olmo 3 1025 7B. Apple Silicon uses unified memory, so the model shares RAM with the system. At Q4_K_M you need at least 5.8 GB of usable unified memory (RAM minus macOS overhead).

Can I run Olmo 3 1025 7B locally?

Yes — Olmo 3 1025 7B can run locally on consumer hardware. Popular tools include Ollama, LM Studio, and llama.cpp.

How fast is Olmo 3 1025 7B?

At Q4_K_M, Olmo 3 1025 7B can reach ~835 tok/s on AMD Instinct MI350X. On NVIDIA GeForce RTX 4090: ~114 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 ÷ 5.75 × 0.65 = ~904 tok/s

Estimated speed at Q4_K_M (5.8 GB)

NVIDIA B200

~904 tok/s

NVIDIA GeForce RTX 4090

~114 tok/s

NVIDIA B300

~904 tok/s

AMD Instinct MI350X

~835 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 Olmo 3 1025 7B?

At Q4_K_M, the download is about 4.38 GB. The full-precision BF16 version is 14.60 GB. The smallest option (IQ2_XXS) is 2.01 GB.

Which GPUs can run Olmo 3 1025 7B?

52 consumer GPUs can run Olmo 3 1025 7B at Q4_K_M. Top options include AMD Radeon RX 6700 XT, AMD Radeon RX 6800, AMD Radeon RX 6800 XT, AMD Radeon RX 7600. 40 GPUs have plenty of headroom for comfortable inference.

Which devices can run Olmo 3 1025 7B?

151 devices with unified memory can run Olmo 3 1025 7B at Q4_K_M (5.8 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.