Segmind Release Notes (October 12–18, 2026)

1. New Image Models

• What's new: Several image models are now live on the API: FLUX 3 Klein (flux-3-klein-text-to-image and flux-3-klein-image-edit), Krea 3 (krea-3-text-to-image and krea-3-image-edit), and Qwen-Image 2.1 Turbo (qwen-image-2.1-turbo) for faster image generation. flux-3-text-to-image and flux-3-image-edit also now offer a 1.5k resolution tier.
• Impact: Generate and edit images with the latest FLUX, Krea and Qwen models, and pick a faster turbo option when speed matters.
• Released: October 13–17, 2026

2. New Video and Avatar Models

• What's new: PixVerse V6 supports transition, extend and reference-to-video. MiniMax H3 Max (minimax-h3-max) offers text-to-video and image-to-video at 480P, 768P and 1080P, and 1080P is now available on minimax-h3-max-turbo. OmniHuman 2.0 (omnihuman-2.0) turns one portrait and 4–300 seconds of audio into a video of that person speaking or singing, with the video length matching your audio. Two new Pruna models join the P-Video 2 Pro line: p-video-2-pro-avatar for talking avatars and p-video-2-pro-reference for reference-to-video, which has an optional keep_audio flag to keep your own audio as the soundtrack.
• Impact: Create talking-avatar videos from a single photo, extend or transition between clips, steer videos with reference media.
• Released: October 13–19, 2026

3. Gemini 4 Argon and Claude Tool Calling

• What's new: Gemini 4 Argon (gemini-4-argon), Google's flagship reasoning model, is now available for text tasks and supports longer videos of around three hours. Claude models now support tool calling on the API, including tools, tool_choice and system prompts, and system messages in OpenAI-style requests are now handled correctly. File and image inputs sent to Claude models are converted automatically, and unsupported file inputs on other chat models now return clear errors.
• Impact: Build agents and function-calling workflows on Claude using the request format you already use, and analyze much longer videos with Gemini.
• Released: October 14–17, 2026

4. Voice Studio

• What's new: Voice Studio is live with elevenlabs-v4-tts, HeyGen voice text-to-speech and voice cloning, and a new /v1/voices endpoint. Voice pickers in the console show voices by provider, and you can delete your own voices. HeyGen voices also work inside Pixelflow.
• Impact: Manage your custom voices in one place and use them across the API, the console and Pixelflow.
• Released: October 12–13, 2026

5. Creator Pages and Posts

• What's new: Creators can now publish public pages at segmind.com/@handle, bundle their work with a one-page composer, and write posts with embedded blocks. A new blog-style editor supports rich text, a / menu, media and prompt widgets, and Markdown mode. A new skill block lets creators publish a SKILL.md that readers can copy, download or install.
• Impact: Share your Pixelflow templates, prompts and skills with the community from a page of your own.
• Released: October 17–19, 2026

6. Pixelflow Improvements

• What's new: Model pickers now show a Your Models row with your dedicated endpoints and any models enabled for your account or team. Output nodes show media details such as format, resolution, size and duration in a compact info pill. Annotate and mask image inputs now connect correctly as single files. Templates with large card images now publish reliably, and API runs on the workflow worker send their PIXELFLOW_API_RUN webhooks again. The Pixi assistant can now edit your open workflow live and sees the canvas with each message.
• Impact: Find restricted models faster, inspect outputs at a glance, and publish and automate workflows with fewer interruptions.
• Released: October 12–18, 2026

7. Account, Security and Billing

• What's new: You can now turn on optional two-factor authentication (TOTP) in your security settings, and a new password-reset page is available. Login links and password-reset links are now single-use. Creating a dedicated endpoint now requires an hour's worth of balance, and resumed endpoints restore your last chosen configuration. Inviting a teammate from a Personal account now creates your team automatically. Creating your first API key now always shows the full key. The recharge dialog explains why checkout could not start, and postpaid emails include your receipt link and invoice number again. Predictions that complete without delivering a result are no longer charged.
• Impact: Better protection for your account, clearer billing, and smoother team and endpoint setup.
• Released: October 13–19, 2026

8. Console, Support and Discovery

• What's new: The contact page now lets you choose Talk to sales or Get support, so support requests go straight to the support team. The Cloud UI has Status and Source filters and a cleaner layout. Pricing and API tabs now label partner models correctly, and the MCP page lists Grok Build. Generations for Seedance 2.5 drafts reappear in History and Generations.
• Impact: Reach the right team faster and find your generations and models more easily.
• Released: October 13–18, 2026