Quickstart
Run Qwen3.8-27B on one Google Cloud TPU v6e chip by hand, with three gcloud commands from your own machine: create the chip, install and run, delete the chip. Install does the same in two lines, with mashbox tpu up.
You need the Google Cloud CLI signed in to a project with billing and the TPU API turned on. Run on Google Cloud TPU steps 1 to 5 do that from a new Google account.
gcloud compute tpus tpu-vm create mashbox-1 --zone=us-east5-b --accelerator-type=v6e-1 --version=v2-alpha-tpuv6egcloud compute tpus tpu-vm ssh mashbox-1 --zone=us-east5-b --command='python3 -m pip install uv && python3 -m uv tool install --python 3.12 "mashbox[tpu]" && ~/.local/bin/mashbox run qwen3-8-27b --prompt "Write a haiku about the sea."'gcloud compute tpus tpu-vm delete mashbox-1 --zone=us-east5-b --quietThe second line installs uv and mashbox with Python 3.12 (the chip’s own Python is 3.10, too old for JAX on TPU), downloads the 26.94 GB weights file, checks it and prints the answer. It prints an answer like this one (TPU v6e-1):
Salt spray kisses stone,Blue waves roll in endless rhythm,Sun sets on the tide.- Serve it with an OpenAI-compatible API: Qwen3.8-27B, Serve it.
- Use it from Claude Code, Codex or Pi: Use with coding agents.
- Something went wrong: Frequently asked questions.