0G Private Computer has processed more than 250 billion tokens through 17 million requests, as the private artificial intelligence platform expands to 28 models and introduces payments in US dollars.
0G Labs announced the figures in a press release on 18 August, saying the service had also generated 15 million Trusted Execution Environment (TEE) attestation proofs since its launch on 27 April.
Private Computer initially offered seven models. That number has now increased to 28, available through a web console and an application programming interface compatible with OpenAI’s API. Users can access text, image, audio and video generation tools through a single account.
The company said Private Computer does not retain user data. Prompts are processed in memory, deleted after use and excluded from model-training datasets. Usage statistics and attestation activity are published on a real-time public dashboard.
Dollar payments introduced
USD Mode is now generally available, allowing users to hold account balances in US dollars and add funds using a bank card through Stripe.
The updated service also includes email sign-in and a guided process for making a first deposit. Previously, users interacting with decentralised AI infrastructure could have needed to create a crypto wallet, purchase a supported token and complete an on-chain transaction.
Under the new process outlined by 0G Labs, users can register with an email address and begin using Private Computer without buying a token.
For customers in the United States, the principal change is the option to pay in dollars through a familiar card-based checkout. 0G Labs did not announce a new US-specific token, investment product or securities offering as part of the update.
“Privacy should not require a crypto wallet,” 0G Labs co-founder and chief executive Michael Heinrich said.
Users can choose a trust tier for each request. Verifiable execution is available on 20 of the 28 models, while payments can still be made in dollars when a selected model uses the platform’s decentralised verification infrastructure.
Trusted Execution Environments are protected areas within computer hardware that isolate data and code while a task is being carried out. According to 0G Labs, Private Computer uses the technology to process prompts without exposing them outside the protected environment.
Of the 20 models with verifiable execution, 15 use TeeTLS for verified routing. A further five use TeeML, which keeps prompts inside a hardware enclave while the model produces its response.
Each model page includes privacy labels showing users what type of execution is available before they submit a request. Eight of the 28 models do not currently offer the same verifiable-execution support, based on figures supplied by the company.
TEE-based computing is also being tested elsewhere in the cryptocurrency and blockchain sector. A March report on Origin detailed an $8m funding round for a modular AI network designed to separate demanding model workloads from blockchain-based verification.
In April, crypto.news reported on 0G’s Qianwen integration, which connected Alibaba’s large language model family with applications and AI agents built on 0G infrastructure. The integration was intended to allow developers to use Qianwen models in on-chain applications instead of relying solely on conventional AI platforms.
Private Computer presents the same underlying infrastructure in a consumer-facing format. Rather than requiring users to manage the blockchain layer directly, it provides access through an account dashboard and a standard API used by many AI developers.
Video model added
Hailuo 3, a video model developed by MiniMax, was added to Private Computer on 3 August, according to the release. Through the browser interface, it supports both text-to-video generation and first-frame image-to-video generation.
Its inclusion means customers can use video, image, text and audio models without opening separate accounts for each format. The number of models available through Private Computer has quadrupled from seven at launch to 28, while 0G Labs said it had reached 15,000 users during the same period.
Developers can also access the models through the OpenAI-compatible API. That compatibility means applications already built around the API format can connect to 0G’s service without requiring a completely different request structure.
0G Labs said users can select both a model and a verification level for individual requests. The public dashboard records aggregate activity, including the reported 250 billion tokens, 17 million requests and 15 million TEE proofs.
Private Computer serves as an access point to other parts of 0G’s network, including decentralised computing, distributed storage, data availability services and an Ethereum Virtual Machine-compatible blockchain.
The blockchain, known as Aristotle Mainnet, launched alongside the 0G token in September 2025. A report on its listing on Kraken described 0G as a modular layer-one network designed to support decentralised AI applications.
That September report said the project had raised more than $350m, including a $40m seed round and a $250m token purchase commitment secured by the 0G Foundation. Investors named in the earlier funding included Hack VC, OKX Ventures, Delphi Ventures, Samsung Next and Animoca Brands.
In its latest release, 0G Labs said its ecosystem had now raised more than $360m and was working with more than 100 partners. The company listed Chainlink, Google Cloud and Alibaba Cloud among the organisations connected to the ecosystem.
