xAI is set to release a new multi-modal version of its Grok model next week, which will include vision capabilities. The updated Grok model will also feature a 128k context length, function calling, and system prompts. Additionally, Grok Beta with multi-modality will be available soon, and the API roadmap includes a vision model release later this week, along with other enhancements such as an API playground, better documentation, auto-upgrade of rate limits, Tokenizer API, and SDKs. Planned features with release dates to be determined include prompt caching, batch inference, fine tuning, and real-time information.