OpenRouter listed a free anonymous AI model named Ox Alpha on August 20–21, 2026, carrying a 1,048,576-token context window and a stated design focus on coding, sustained agentic work, and production workflows. The provider is identified only as “stealth” — no company, no lab, no benchmark numbers.
Ox Alpha accepts text, images, and video as input and returns text. OpenRouter lists a maximum output of 131,072 tokens and prices both prompt and completion tokens at $0 for a preview window the OpenCode team describes as roughly one week. OpenRouter’s listing states that the provider retains prompts but does not use them for training. The OpenCode team confirmed “near unlimited usage, zero data retention” for the free window. A Hacker News discussion of the launch reached 178 points on August 21 and characterises the model as built for long-horizon software engineering, complex reasoning, and workflows that combine text with visual context.
“This model is stealth — developed and operated by a third-party provider who has chosen to remain anonymous during this preview.” — OpenRouter listing
Why a $0 Model With 1M Context Changes the Experimentation Math
Ox Alpha removes the token bill from long-context agentic testing, which is the single largest cost line in codebase-scale AI experiments. A 1M-token window holds an entire mid-size repository, a full quarter of support tickets, or a multi-hundred-page contract set in one prompt — work that costs $3–15 per million tokens on Claude Sonnet 5 and GPT-5.6 Lun, and $0 on Ox Alpha during the preview.
The workloads that benefit are the ones businesses already run against paid frontier APIs: repository-wide refactors, multi-step data processing, document-set analysis, and CI failure triage. Teams comparing those jobs across providers can measure the same benchmark-backed options in our guide to the best AI agents for business tasks, then run the identical prompts through Ox Alpha at zero marginal cost for the length of the window.
What Nobody Can Verify About Ox Alpha Yet
No official benchmark results exist for Ox Alpha as of August 21, 2026. The provider has released no evaluation numbers, no model card, and no architecture detail, so every performance comparison circulating this week is anecdotal. Three further facts remain unverified: the builder’s identity, whether the model stays free after the preview, and how much of the 1,048,576-token window is usable in practice rather than merely advertised.
The data-retention position carries the same caveat. OpenRouter states that prompts reach the provider and are excluded from training, but an anonymous operator offers no contract, no jurisdiction, and no audit path behind that statement. Businesses evaluating Ox Alpha alongside the vendor-backed AI tools for business they already license should treat it as an unaccountable endpoint and keep regulated, customer, and proprietary data out of it.
The Stealth-Drop Pattern Points at a Chinese Lab
Every previous stealth model on OpenRouter was later claimed by a Chinese AI company, which is why the community reads Ox Alpha the same way. The four precedents are Zhipu AI (GLM-5), Xiaomi (MiMo-V2-Pro), Ant Group (Lingxi Ling-2.6-flash), and Meituan (LongCat-2.0). Xiaomi’s MiMo team is the most widely-cited candidate for Ox Alpha, and that attribution is community speculation — the provider has confirmed nothing.
The mechanism explains the secrecy. OpenRouter’s stealth program gives a lab real production traffic and real user behaviour data before a formal launch, without the PR scrutiny or competitive signalling that a named release triggers. The precedent set by GLM-5 and MiMo-V2-Pro suggests the reveal follows within weeks of the preview, not months.
For Context: Our Coverage of OpenRouter and Chinese Frontier Models
- Stripe’s $7B OpenRouter acquisition — who now owns the router that hosts every stealth drop, and what that means for your AI stack.
- GLM-5.2 review — Zhipu AI’s 1M-context model, and the lab behind the first OpenRouter stealth model to be unmasked.
- GLM-5.3 coding results — the current open-source coding leader Ox Alpha will be measured against once benchmarks appear.
- Qwen3.8-27B open weights — the free-to-run alternative that carries a named lab and a published licence.
- DeepSeek V4-Pro pricing — what a Chinese frontier model charges once it exits preview, and the rate Ox Alpha’s free window will be judged against.
Ox Alpha is worth an afternoon of testing and none of your production traffic. A free 1M-token window is a real offer for benchmarking your own agentic workloads this week, and a model with no name attached to it has no support contract, no pricing commitment, and no accountable party if the endpoint disappears on day eight. Run your evaluation prompts through it, record the results, and keep customer and proprietary data out. If the OpenRouter stealth pattern holds, the lab reveals itself within weeks — that is the moment to decide whether Ox Alpha belongs in your stack.
