xAI
AvailablexAI has 11 models tracked on Modelglass across 5 verticals (Language, Video, Audio, Coding, GDPval), most recently added 2026-10.
xAI has 4 language models (1 current-gen, 2 previous-gen, 1 retired), 1 video model (1 current-gen), 2 audio models (1 current-gen, 1 deprecated), 2 coding models, and 2 gdpval models on Modelglass, most recently added 2026-10.
Coverage across verticals
1 current-gen · 2 previous-gen · 1 retired
Language pricing page ↗Creator and first-party host of the Grok model family. Standard per-1M-token input/output pricing published at x.ai/api.
1 current-gen
Video pricing page ↗xAI's own first-party API. Grok Imagine offers both video and image generation; video is billed per second of output at a flat rate — no resolution-based pricing tiers on the current docs.x.ai model pages.
1 current-gen · 1 deprecated
Audio pricing page ↗xAI (a subsidiary of SpaceX as of April 2026) offers the Grok Voice Think Fast line — speech-to-speech conversational voice models accessible via the xAI API's Voice API, billed per minute of audio. Pay-per-use, no subscription required. Distinct from xAI's text-generation (Grok 4.5) and image/video (Grok Imagine) product lines, which are out of scope for this audio-only registry.
See how xAI's models fit into a cost-aware routing setup
See how →Release cadence
11 models tracked across all verticals. Earliest release 2025-02, most recent 2026-10.
Derived from Modelglass's per-vertical provider registries: the image, LLM, video and audio provider records, plus the science, coding, agentic and GDPval benchmark registries, where a provider is identified by the prefix before "/" in each model's model_id rather than a registry record of its own. Where the same company uses a different provider slug or model_id prefix across verticals (e.g. Google DeepMind vs. Google Cloud, or Qwen's "alibaba" model_id prefix in the coding registry vs. its "qwen" provider slug elsewhere), coverage is merged onto a single profile per an explicit, human-reviewed alias list. Everything else reflects exact-slug or exact-prefix matches only.