Arcee AIAfmoeForCausalLM

Trinity Mini — Hardware Requirements & GPU Compatibility

ChatFunctions

Trinity Mini is a 26.1B-parameter open language model from Arcee AI. It supports a context window of up to 131,072 tokens. At Q4_K_M it needs about 16.04 GB of VRAM — see which GPUs and Macs can run it below.

26.1B params · MoE3.5B active131K contextOtherDec 202512.1K downloads206 likes48.5K 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.

S

Runs great on a 24 GB GPU

Runs from 12 GB (Q2_K · 16K)Best on 24 GB (Q4_K_M · ~199 tok/s · 128K)Macs from 16 GB (Q2_K · ~58 tok/s)

Other

Reference GPU · RTX 4090 24 GB

Estimates, ±20% · how we estimate

Run Trinity Mini

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

Recommended

Ollama

one command · needs Ollama installed

$ollama run hf.co/MaziyarPanahi/Trinity-Mini-GGUF:Q4_K_M

Q4_K_M fits fully on the RTX 4090 · 7.5 GB free · 16K context

Download: Trinity-Mini.Q4_K_M.gguf15.82 GBMaziyarPanahi

install, OS, other quants

Other ways

each opens step-by-step

Run in cloud

Fits on RTX A5000 (24 GB) (7 GB headroom) · Q4_K_M

Generation speed: ~166 tok/s; generation speed
Cost per 1M output tokens: $0.27; per 1M output tokens

Deployon RunPodBest optionfrom $0.16/hr

Compare GPUs →

or

How Much VRAM Does Trinity Mini Need?

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

Quantization	VRAM	+32K	+128K	File	Quality	On RTX 4090 (reference)
	11.5 GB	12.5 GB	15.7 GB	9.7 GB	2-bit · K-quant — 2-bit quantization with K-quant improvements	Fits: 128K context · ~245 tok/s
	13.1 GB	14.1 GB	17.3 GB	12.5 GB	3-bit medium — 3-bit medium quantization	Fits: 128K context · ~226 tok/s
	13.8 GB	14.8 GB	18 GB	13.6 GB	3-bit large — 3-bit large quantization	Fits: 128K context · ~220 tok/s
	16.0 GB	17.1 GB	20.3 GB	15.8 GB	4-bit medium — 4-bit medium quantization — most popular sweet spot	Fits: 128K context · ~199 tok/s
	19.0 GB	20 GB	23.2 GB	18.6 GB	5-bit medium — 5-bit medium quantization — good quality/size tradeoff	Fits: 128K context · ~178 tok/s
	21.9 GB	22.9 GB	26.2 GB	21.5 GB	6-bit, very good — 6-bit quantization, very good quality	Fits: 64K context · ~161 tok/s
	26.5 GB	27.5 GB	30.7 GB	~26.1 GB	8-bit, near-lossless — 8-bit quantization, near-lossless	Does not fit: needs 27 GB
	52.6 GB	53.6 GB	56.9 GB	~52.3 GB	16-bit · bfloat — Brain floating point 16 — preferred for training	Does not fit: needs 53.1 GB

est.calculated, no published file yet

Selected Q4_K_M

4-bit medium quantization

16.5GBat 16K context

Weights: 15.7 GB
KV cache: 0.5 GB
Overhead: 0.3 GB

Fits RTX 4090 (reference) · 7.5 GB free · ~199 tok/s

≥50 tok/s needs ≥180 GB/s · RTX 5090 ~278, RTX 3090 Ti ~199

Which Hardware Can Run Trinity Mini?

Q4_K_M · 16.0 GB

Trinity Mini (Q4_K_M) requires 16.0 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 21+ GB is recommended. Using the full 131K context window can add up to 4.2 GB, bringing total usage to 20.3 GB. 8 GPUs can run it, including NVIDIA GeForce RTX 5090, NVIDIA GeForce RTX 3090 Ti. 122 devices with unified memory can run Trinity Mini, including Mac Studio (M3 Ultra, 256GB), Mac Studio (M3 Ultra, 512GB), Mac Mini M5 Pro (24 GB).

By VRAM tier · typical GPUs and Macs at Q4_K_M

Typical hardwareQuantVRAMtok/sContext that fits

8 GB · RTX 4060Doesn't fit — smallest quant needs 11.9 GB8 GB tier →

12 GB · RTX 3060 12GBQ2_K · 11.9 GB · ~121 tok/s · 16K11.9~12116K12 GB tier →

16 GB · RTX 5060 Ti 16GBQ3_K_L · 14.2 GB · ~124 tok/s · 64K14.2~12464K16 GB tier →

24 GB · RTX 4090referenceQ4_K_M · 16.5 GB · ~199 tok/s · 128K16.5~199128K24 GB tier →

Mac 32 GB · Mac Mini M4Q4_K_M · 16.5 GB · ~31 tok/s · 128K16.5~31128KMacs →

Mac 64 GB · Mac Studio M4 MaxQ4_K_M · 16.5 GB · ~84 tok/s · 128K16.5~84128KMacs →

Alternatives to Trinity Mini

Similar models

6tok/s on RTX 4090

Model	Score	VRAM	tok/s	Why it is similar
Qwen3.8 27BAlibaba	122.1	17.9 GB	~38	same size
Qwen3.6 27BAlibaba	111.2	17.9 GB	~38	same size
Qwen3.5 27BAlibaba	103.8	17.9 GB	~38	same size
Gemma 3 27B ITGoogle	79.4	21.7 GB	~36	same size
Gemma 2 27B ITGoogle	72.7	19.5 GB	~37	same size
Gemma 4 26B A4BGoogle	—	16.9 GB	~153	same size

Specifications

Publisher: Arcee AI
Parameters: 26.1B
Active parameters: 3.5B per token
Architecture: AfmoeForCausalLM
Context Length: 131,072 tokens
Vocabulary Size: 200,192
Release Date: 2025-12-01
License: Other

Resources

Hugging Face model card (opens in a new tab)

Frequently Asked Questions

How much VRAM does Trinity Mini need?

Trinity Mini requires 16.0 GB of VRAM at Q4_K_M, or 52.6 GB at BF16. Full 131K context adds up to 4.2 GB (20.3 GB total).

VRAM = Weights + KV Cache + Overhead

Weights = 26.1B × 4.8 bits ÷ 8 = 15.7 GB

KV Cache + Overhead ≈ 0.3 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 ≈ 4.6 GB (at full 131K context)

VRAM usage by quantization

Q4_K_M

16.0 GB

Q4_K_M + full context

20.3 GB

Learn more about VRAM estimation →

Can NVIDIA GeForce RTX 4090 run Trinity Mini?

Yes, at Q6_K (21.9 GB) or lower. Higher quantizations like Q8_0 (26.5 GB) exceed the NVIDIA GeForce RTX 4090's 24 GB.

What's the best quantization for Trinity Mini?

For Trinity Mini, Q4_K_M (16.0 GB) offers the best balance of quality and VRAM usage. Q5_K_M (19.0 GB) provides better quality if you have the VRAM. The smallest option is Q2_K at 11.5 GB.

VRAM requirement by quantization

Q2_K

11.5 GB~75%

Q3_K_L

13.8 GB~86%

Q4_K_M ★

16.0 GB~89%

Q5_K_M

19.0 GB~92%

Q6_K

21.9 GB~95%

BF16

52.6 GB~100%

★ Recommended — best balance of quality and VRAM usage.

Learn more about quantization →

Can I run Trinity Mini on a Mac?

Yes — Mac Mini M2 Pro (16 GB) and 133 other Macs can run Trinity Mini, the smallest at Q2_K. Apple Silicon uses unified memory, so the model shares RAM with the system. At Q4_K_M you need at least 16.0 GB of usable unified memory (RAM minus macOS overhead), from Mac Mini M5 Pro (24 GB).

Can I run Trinity Mini locally?

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

How fast is Trinity Mini?

At Q4_K_M, Trinity Mini can reach ~146 tok/s on AMD Instinct MI350X. On NVIDIA GeForce RTX 4090: ~199 tok/s. Speed depends mainly on GPU memory bandwidth. Real-world results typically within ±20%.

tok/s = 1000 ÷ (active GB ÷ (bandwidth GB/s × efficiency) × 1000 + layers × routing ms)

Mixture-of-Experts: just the active experts get read for each token, along with a fixed routing cost per layer.

Example: NVIDIA B200 → 2.1 GB active ÷ (8000 × 0.65) = 0.41 ms, plus 32 layers × 0.055 ms = 1.76 ms, so 1000 ÷ 2.17 ms = ~461 tok/s

Estimated speed at Q4_K_M (16.0 GB)

NVIDIA B200

~461 tok/s

NVIDIA GeForce RTX 4090

~199 tok/s

NVIDIA B300

~461 tok/s

NVIDIA GH200 Grace Hopper Superchip

~412 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 Trinity Mini?

At Q4_K_M, the download is about 15.67 GB. The full-precision BF16 version is 52.25 GB. The smallest option (Q2_K) is 11.10 GB.

Which GPUs can run Trinity Mini?

8 consumer GPUs can run Trinity Mini at Q4_K_M. Top options include NVIDIA GeForce RTX 5090, AMD Radeon RX 7900 XT, AMD Radeon RX 7900 XTX. 1 GPU has plenty of headroom for comfortable inference.

Which devices can run Trinity Mini?

122 devices with unified memory can run Trinity Mini at Q4_K_M (16.0 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.

Run Trinity Mini

16.5 GB VRAM at 16K context · fits the RTX 4090 (7.5 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/MaziyarPanahi/Trinity-Mini-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/MaziyarPanahi/Trinity-Mini-GGUF, then pick Q4_K_M. Or use the CLI (run LM Studio once first):
$lms get https://huggingface.co/MaziyarPanahi/Trinity-Mini-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 MaziyarPanahi/Trinity-Mini-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/MaziyarPanahi/Trinity-Mini-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 MaziyarPanahi/Trinity-Mini-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 arcee-ai/Trinity-Mini --max-model-len 16384
Runs the original weights, not the Q4_K_M GGUF; needs ~57 GB of GPU memory.

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.
$MaziyarPanahi/Trinity-Mini-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 A5000 (24 GB)CheapestCheapest7 GB free	~166 tok/s	$0.27	Deploy on RunPodfrom $0.16/hr (opens in new tab)	24 GB · 7 GB free · ~166 tok/s · $0.27 / 1M tokens
RTX 4000 Ada (20 GB)3 GB free	~92 tok/s	$0.60	Deploy on RunPodfrom $0.20/hr (opens in new tab)	20 GB · 3 GB free · ~92 tok/s · $0.60 / 1M tokens
RTX A5000 (24 GB)7 GB free	~166 tok/s	$0.36	Deploy on Vast.aifrom $0.21/hr (opens in new tab)	24 GB · 7 GB free · ~166 tok/s · $0.36 / 1M tokens
RTX 3090 (24 GB)7 GB free	~190 tok/s	$0.32	Deploy on RunPodfrom $0.22/hr (opens in new tab)	24 GB · 7 GB free · ~190 tok/s · $0.32 / 1M tokens
TITAN RTX (24 GB)7 GB free	~151 tok/s	$0.45	Deploy on Vast.aifrom $0.24/hr (opens in new tab)	24 GB · 7 GB free · ~151 tok/s · $0.45 / 1M tokens
RTX 3090 (24 GB)7 GB free	~190 tok/s	$0.39	Deploy on Vast.aifrom $0.27/hr (opens in new tab)	24 GB · 7 GB free · ~190 tok/s · $0.39 / 1M tokens
RTX PRO 4000 (24 GB)7 GB free	~151 tok/s	$0.59	Deploy on Vast.aifrom $0.32/hr (opens in new tab)	24 GB · 7 GB free · ~151 tok/s · $0.59 / 1M tokens
RTX A6000 (48 GB)31 GB free	~166 tok/s	$0.55	Deploy on RunPodfrom $0.33/hr (opens in new tab)	48 GB · 31 GB free · ~166 tok/s · $0.55 / 1M tokens

Compare GPUs →