Ox Alpha Speculated to be Microsoft's Unreleased MAI-2 Frontier Model
inductionheads · x · 2026-08-22
Based on tokenizer analysis, the stealth model Ox Alpha is suspected to be using the cl100kbase tokenizer. This rules out major labs like OpenAI, Google, Anthropic, and xAI, pointing instead to Microsoft's Phi/MAI lineage or IBM Granite. The model is likely Microsoft's unreleased "MAI-2". It is currently free to try, offering 1M context, multimodality, and zero data retention.
Related event: Forensics Point to Zhipu's Unreleased GLM Behind Mystery Model Ox Alpha(31 posts)→
More from Models
- Unredacted Anthropic filing exposes internal data-quality debate months before Opus 3 cutoff — zacharynado · 2026-10-15
- Google says teams across the company are using Gemini 4 Argon on its hardest engineering problems — CurieuxExplorer · 2026-10-15
- Perplexity's new embedding models: index with a 9B model, search with 0.6B — antoine_chaffin · 2026-10-15
- 797GB of n-gram counts in CPU RAM let 1B models match 57% more compute — PangWeiKoh · 2026-10-15
- Leak: Moonshot's Kimi flagship 'x4' done with pre-training, K3.1 in internal testing — lyraxana · 2026-10-15
- Unsloth ships GGUF quants for Nemotron 3.5: Bolt-74B runs in 24GB RAM — danielhanchen · 2026-10-15