v15.2.0•
Expanded Model Access & Free Tier Additions
SreeAuthor
This update democratizes access to premium image generation models for Basic tier users and introduces experimental models for Free tier users, while bringing the advanced Grok-4 model to all subscription tiers.
New Features
4
- Grok-4 Model Added: Introduced
grok-4-0709advanced reasoning model with function calling support, available across all tiers (Free, Basic, Pro, Ultra) with 256K context window, 16K max output, 2-5s latency, and 97.5% uptime. - Basic Tier Image Generation Expansion: Added professional-grade image generation models to Basic plan via Provider 3:
imagen-3.0-generate-002- Google's high-fidelity text-to-image generationFLUX.1-dev- 12B parameter text-to-image modelflux-kontext-dev- Advanced image editing capabilities
- Free Tier Experimental Models: Added cutting-edge experimental models to Provider 1 for Free tier users:
- Enhanced Vision Support: Expanded model support to include a new flagship reasoning model and enabled vision capabilities for select providers.
Improvements
2
- Provider 7 Multimodal Enhancement: Added full vision capabilities for Claude models with HTTP URL and data URI image processing support, asynchronous image downloads with base64 encoding, and improved image handling pipeline.
- Provider 7 Async Integration: Updated adapter calls with proper async/await patterns and correct HTTP client passing for improved performance.
Bug Fixes
1
- Provider 7 Token Management Fix: Fixed reasoning budget calculations in max tokens logic to ensure proper token allocation and prevent API errors from incorrect token limits.
Deprecations
1
- Deprecated Image Models: Removed Provider 3 image generation models:
shuttle-3.1-aesthetic,shuttle-3-diffusion,shuttle-jaguar.