Governed · Compressed Run a bigger model in your VPC.
A model that does not fit on the card you have costs a second card, a smaller model, or a cloud bill. Tileward Models fits it.
Works with the clients you already run
27 have a recipe to paste, and anything else that speaks MCP or the OpenAI wire format works too.
A compressed model has three jobs. Two are measured, one is where it runs.
Fit on the hardware you have, stay the model you chose, and run wherever you run it. The first two are measured against the publisher’s own release.
Fits the card you have
smaller (1.3–4.5×) than the weights as published, across the ten models in the library that we have compressed. A mixture of experts lands 24% to 78% smaller (1.3× to 4.5×); a dense model lands 64% smaller (2.8×).
Still the model you chose
on MMLU, written down before each run as the most a build may fall behind the original. GLM-5.3, Kimi K3 and Kimi K3.1 Pro each stayed inside it.
Runs where you run it
places it runs unchanged: our hosted API, your VPC, or air-gapped.
What does fitting the card save?
A model that needs fewer cards costs less to run. Pick a model and the cards you have, and compare the weights as published with Tileward’s.
Estimate it by hand: divide the model’s weights by what one card holds after the VRAM you set aside for context, round up, and multiply by the copies you run and what a card costs an hour. Do it once with the weights as published and once with Tileward’s, sizes on the catalogue below.
| Line | As published | With Tileward |
|---|
Every model served today, in one list.
Each one is listed with the size, context window, price and the id to call it by that the library gives it.
Tileward Qwen3.8-27BVoxtral RealtimeTileward Whisper large-v3Tileward Whisper large-v3-turbo
More are ready to serve, hosted or on your hardware. Talk to an engineer
Browse the model library →Your hardware, your VPC, the same governance.
24% to 78% smaller (1.3× to 4.5×)
Mixtures of experts land 24% to 78% smaller than the weights they were compressed from (1.3× to 4.5×). A dense model lands 64% smaller (2.8×).
OpenAI-compatible
Change the base URL and your existing client works.
Governed by the same tiles
The 215 Tileward Governance tiles apply to the self-hosted model exactly as they do to the hosted API, with audit records retained locally.
client.chat.completions.create(model="Tileward-Qwen3.8-27b", messages=[...])
# same call self-hosted, only the base_url changes
Three ways to run it
- Change the base URL
- Default tile pack
- Zero infrastructure
- Single node, one GPU
- Tiles and policies of your own
- Data never leaves
- Fully offline
- Runs on your hardware
- Local audit retention