What's New

A running summary of what shipped on Kolbo.AI, most recent first. For exact model rates see Models & Pricing.

June 2026

New models

LTX 2.3 Quality by Lightricks: high-end cinematic video from text, image animation, and Ingredients mode (up to 9 reference images), with native audio, up to 4K, and 5 to 20 second clips
HappyHorse 1.1 by Alibaba: text-to-video, image-to-video, and reference-to-video (up to 9 references), with native audio, 720p and 1080p, and lower 1080p cost
Luma Ray 3.2: flagship cinematic text-to-video plus image-to-video, video-to-video, first-last-frame, and reframe modes
Sync-3 lipsync: most advanced lip-sync model, with a new Sync Mode to control how mismatched audio and video line up

Chat & AI

Voice Agent: start a live voice call with your AI assistant, speak naturally, get spoken replies in real time, and share your camera or screen mid-call
Smarter Prompt Copilot: wires dropped images straight into your prompt, keeps the model you picked, understands @Visual-DNA and #moodboard references, and points you to the right Learning Hub lesson
Web Search on any model: your chosen model answers from live results with numbered inline citations and a sources list, charged only when a search runs

Projects & workflow

Project Knowledge: give each project a living memory from uploaded files, links, notes, YouTube videos, and Google Docs, with an auto-maintained Project Profile
Project Profile flows into Chat, the help assistant, and Smart Select, so answers and model picks are grounded in your actual project
Report offensive content: flag any generated result in one tap for team review

Plans & pricing

Refreshed monthly plans with more credits and new Professional 2x and 3x tiers for heavy users. See Subscription & Billing
Bigger credit packs with explicit +10% bonus credits, and cancel a scheduled downgrade by re-upgrading

New models

Grok 4.3, xAI's flagship, in the chat model picker with a 1M-token context window
Gemini Omni Video and Gemini Omni Flash: multimodal video that can pin character voices and audio, 720p to 4K
Grok Imagine Quality for text-to-image and edit, with selectable 1K and 2K
Flux Pro Erase for cleaner removals, plus Flux2 and Flex Max image editing
Midjourney V8 and Niji V7, Topaz Labs video upscaling to 4K, and Ideogram background remove and replace
30 new Gemini text-to-speech voices with a Voice direction picker and Accent selector

Image & creative

Storyboard Grid Creator under Creative Director, with up to 16K export, captions, and Open in AI Board
Attach reference video and a voice sample when creating a Visual DNA, plus a new Erase mode in Inpaint
Auto-enhance when a preset or moodboard is attached, and aspect ratio chips on every generation

Developer API

Route any generation into a specific project with an optional project_id, plus a new GET /v1/projects endpoint
Full media library endpoints, quick-share for HTML artifacts, and richer model metadata in list_models

Chat & workflow

Presentations, landing pages, and interactive HTML artifacts generated directly in Chat
Live Generation ETA, Visual DNA Recently Used, and move sessions between projects

Smart Generation Routing: a circuit breaker that reroutes around temporarily unavailable models
Faster Creative Director with batched multi-scene generation
New music provider Sonauto (v3 variable-length songs, v2 with extend support) at 15 credits per song
Clearer usage history breakdown across generation types

New models

ByteDance Dola Seed 2.0 Mini multimodal analysis and SeeDream 5.0 Lite image generation
Training Lab expansion across multiple base models (flux.2, qwen, krea, wan, z-image)

Video & audio

Video Editor batch operations, new 21:9 and 4:5 aspect ratios, and Cinema and Full Screen modes
Motion Graphics with synchronized music and sound, and real-time editing
Speaker detection in transcriptions

Workflow & performance

Smart Project Memory, PDF and text context upload, and project files across chats
Instant language switching, a real-time credit counter, and faster auto-model performance

New ByteDance video upscale model with speed improvements
Creative Director optimizations and a Zoom Out feature for Reframe
Media Library custom folders, move items between folders and projects, and session moves
Desktop Format Factory integration for media conversion and trimming