TT-QuietBox ® 2

Run and build on top of models

Run & build

6 chapters · 48 min

Tinker with models and the software & hardware stack

Tinker

6 chapters · 59 min

Customize your workstation and make it truly your own

Customize

5 chapters · 36 min

Flow-matching transformer for photorealistic text-to-image with strong prompt adherence....

TT-QuietBox^® 2

Blackhole® · 4 chips · liquid-cooled · $9,999

Product page ↗

AI Accelerator

Chips4× Blackhole® (2× p300c cards)

Tensix cores480 total (120 per chip)

AI clock1.35 GHz

SRAM720 MB total (180 MB/chip)

DRAM128 GB GDDR6 total (32 GB/chip)

Bandwidth1,024 GB/s per card

Chip interconnectWarp400 (2× per card, Samtec ARP6)

Host System

CPUAMD Ryzen 7 9700X (8c, 3.8 GHz)

System RAM256 GB DDR5-5600

Storage4 TB NVMe PCIe Gen5

MotherboardASRock B850M-C

OSUbuntu 24.04.3 LTS

PSU1,600 W

Peak draw~1,500 W

Connectivity

Display1× HDMI

USB8× (2× Gen2, 2× Gen1, 4× USB 2.0)

LANGigabit Ethernet

WiFiWi-Fi 6 (802.11ax) dual-band

Bluetooth5.3

Physical

Dimensions9.1 × 17.8 × 15.6 in

Weight20 kg / 44 lbs

CoolingLiquid (38 dBA max)

Operating temp10–35°C (50–95°F)

Source: docs.tenstorrent.com — TT-QuietBox 2 Specifications

What's Running on QB2

Live from the Tenstorrent compatibility matrix — models confirmed or in testing on Quietbox 2 / P150 / P300c hardware.

Updated 2026-07-01 07:01 UTC

Supported Text-to-Image

FLUX.1 [dev]

12B

FLUX distilled to 4 steps — full quality, fraction of the compute. Black Forest Labs, 12B.

Supported Text-to-Image

FLUX.1 [schnell]

12B

Multilingual instruction following, tool use, and function calling. Meta, 8B.

Llama-3.1-8B-Instruct

Stronger structured tasks, tool use, and reasoning than prior Llama generations. Meta,...

Llama-3.3-70B-Instruct

70B

meta llama/Llama 3.1 8B Instruct

meta llama/Llama 3.3 70B Instruct

Text-to-video focused on motion quality and temporal coherence. Genmo, 10B.

Supported Text-to-Video

Mochi 1

10B

Preview text-to-video from natural language prompts. Motif, 6B.

Supported Text-to-Video

Motif Vision 6B Preview

MMDiT architecture — strong text rendering and prompt control for image generation....

Supported Text-to-Image

SD-XL 1.0-base

6.6B

Mix-transformer segmentation without positional encoding — accurate at low compute....

Supported Text-to-Speech

SpeechT5 (TTS task)

307M

Causal video transformer for text-to-video with strong motion coherence. Alibaba, 14B.

Supported Text-to-Video

Wan2.2

14B

conditional generation U Net for Conditional Generation Base

cv image seg Ultra Fast Lane Detection TuSimple ResNet18 Backbone

cv image seg VGG19 UNet base

Open GPT-style generation at a practical self-hosted scale. GPT-OSS, 20B.

gpt-oss-20b

20B

tt-metal

128K context and multilingual base with a wide fine-tuning ecosystem. Meta, 8B.

Llama-3.1-8B

tt-metal

mm action prediction OpenVLA OFT Finetuned Libero 10

mm action prediction pi_0 pi0 base

mm image text similarity CLIP Base Patch16

mm image text similarity SigLIP Base Patch16 224

mm visual qa Mistral base

nlp causal lm ALLaM 7B Instruct

nlp causal lm arcee_ai arcee Spark

nlp causal lm Falcon 3 10B Base

nlp causal lm Falcon 3 1B Base

nlp causal lm Falcon 3 3B Base

nlp causal lm Falcon 3 7B Base

nlp causal lm Gemma 1.1 2B IT

nlp causal lm Gemma 1.1 7B IT

nlp causal lm Gemma 2 2B IT

nlp causal lm Gemma 2 9B IT

nlp causal lm GPT 2 Base

nlp causal lm GPT 2 Xl

nlp causal lm Llama 3.1 8B Instruct

nlp causal lm Llama 3.2 1B

nlp causal lm Llama 3.2 3B

nlp causal lm Mistral 7B INSTRUCT v03

nlp causal lm Mistral Ministral 8B Instruct

nlp causal lm olmo_3 3 1025 7b

nlp causal lm olmo_3 3 7b instruct

nlp causal lm olmo_3 3 7b think

nlp causal lm Phi 1 LoRA Phi 1

nlp causal lm Phi 1 Phi 1

nlp causal lm Phi 1.5 Phi 1 5

nlp causal lm Phi 2 Phi 2

nlp causal lm Phi 3 Mini 128K Instruct

nlp causal lm Phi 3 Mini 4K Instruct

nlp causal lm Phi 3 Mini Instruct

nlp causal lm Phi 4 Phi 4

nlp causal lm Qwen 2.5 0.5B

nlp causal lm Qwen 2.5 0.5B Instruct

nlp causal lm Qwen 2.5 1.5B Instruct

nlp causal lm Qwen 2.5 14B Instruct

nlp causal lm Qwen 2.5 7B Instruct

nlp causal lm Qwen 2.5 Coder 0.5B

nlp causal lm Qwen 2.5 Coder 1.5B

nlp causal lm Qwen 2.5 Coder 1.5B Instruct

nlp causal lm Qwen 3 0 6B

nlp causal lm Qwen 3 1 7B

nlp causal lm Qwen 3 14B

nlp causal lm Qwen 3 4B

nlp causal lm Qwen 3 8B

nlp causal lm SOLAR 10_7B 10 7B Instruct v1.0

nlp embed gen BERT emrecan/bert base turkish cased mean nli stsb tr

nlp embed gen Qwen 3 Embedding 4B

nlp embed gen Qwen 3 Embedding 8B

nlp token cls BiLSTM CRF Default

99 languages, 680K hours of training audio — built for robust speech recognition. OpenAI,...

Supported Speech-to-Text

whisper-large-v3

1.5B

Largest Gemma 3 — 140-language reasoning, coding, long context. Google, 27B.

gemma-3-27b-it

27B

Experimental

google/gemma 4 31b it

Toggleable chain-of-thought for on-demand deep reasoning. Alibaba, 32B.

Qwen3-32B

32B

Punches above its weight in reasoning and code. Google, 4B.

gemma-3-4b-it

Lightweight agent backbone for low-latency, resource-constrained deployments. Meta, 3B.

Llama-3.2-3B-Instruct

Reasoning depth on demand without a throughput penalty. Alibaba, 8B.

Qwen3-8B

tiiuae/Falcon3 7B Instruct

Experimental