Skip to content
youbot

◆ Changelog

Updates

New models, improvements and fixes — newest first.

  1. New model

    Coming soon: Kimi K3, Gemini TTS & ElevenLabs V3

    Kimi K3 (Moonshot AI), Gemini 3.1 Flash TTS, Gemini 2.5 Pro Preview TTS and ElevenLabs V3 — preview them on the market now; API access lands as soon as they're live.

  2. Improvement

    Mobile & pricing improvements

    Smoother navigation on phones and tablets, and clearer per-configuration labels and pricing on the Pricing page.

  3. New model

    GPT-5.6 Terra, Luna & Sol

    OpenAI's GPT-5.6 in three price/performance tiers — Terra, Luna and Sol — for coding, reasoning and agentic work, available through the chat API.

  4. New model

    Seedance 2.0 Mini

    ByteDance's cost-efficient video model — text- & image-to-video with consistent characters and cinematic motion.

  5. New model

    Grok 4.5

    xAI's latest frontier model for coding, agentic tasks and knowledge work, available through the chat API.

  6. New model

    Grok 4.3

    xAI's flagship reasoning model with text + image input and both standard and deep-reasoning modes.

  7. New model

    Seedream 5.0 Pro

    ByteDance's multimodal image model for controlled generation, precise editing and multilingual visuals.

  8. New model

    Nano Banana 2 Lite

    Google's Gemini 3.1 Flash-Lite Image — fast 1K image generation and prompt-based editing.

  9. New model

    Claude Sonnet 5

    Anthropic's agentic, multimodal model with configurable effort levels, available through the chat API.

  10. New model

    Veo 3.1 & Veo 3.1 Fast

    Text & image to video with native audio is now available, plus a lower-latency Fast tier for rapid iteration.

  11. New model

    Suno V4.5 Plus

    Generate full songs with vocals from a text prompt, with an instrumental-only option.

  12. API

    Unified task API

    A single createTask / recordInfo pattern across every model family, with normalized result URLs.

  13. New model

    Nano Banana Pro

    Fast, photoreal image generation and editing, with reference-image support.

  14. Improvement

    Automatic refunds

    Credits are now automatically refunded when an upstream job fails before producing a result.

  15. Fix

    More reliable polling

    Resolved an issue where some long-running video tasks reported status late.