Chutes has become one of the upstream providers of SayGM, an LLM gateway built on Bittensor subnet 28. Chutes announced the partnership on its site on 28 October 2026.
SayGM runs its gateway inside an Intel TDX trusted execution environment and offers Claude, GPT, Gemini and a range of open-weight models behind one API key and one base URL. Open-weight models that SayGM serves inside a TEE sit in its confidential tier and carry the -tee suffix on the model ID. Chutes says the models it serves run in hardware-attested TEEs with end-to-end encryption.
According to Chutes, nothing changes for existing users: the key, the endpoint llm.chutes.ai/v1 and the catalog at llm.chutes.ai/v1/models stay as they are. Users who route through SayGM can pick a -tee model, and a request can be pinned to Chutes with a header described in the SayGM documentation.
The Chutes product page lists GPU rental with pay-as-you-go billing in TAO and other payment methods. This news concerns the inference gateway, not the GPU rental offer.