Provider 2 Image Generation Expansion & Size Parameter Support

Comprehensive Model Catalog: 31 specialized image generation models now available across all tiers, providing unprecedented access to diverse artistic styles, photorealistic rendering, and rapid generation capabilities.

• Free Tier Models (18 models): Democratizing creative AI for all users

  • provider-2/dreamlike-photoreal-2.0 - Photorealistic image synthesis
  • provider-2/dall-e-2 - OpenAI's proven generation model
  • provider-2/animagine-xl - Anime-style generation
  • provider-2/animagine-xl-2.0 - Enhanced anime synthesis
  • provider-2/juggernaut-x-hyper - High-speed generation
  • provider-2/sdxl-flash - Ultra-fast SDXL variant
  • provider-2/sdxxxl - Extended SDXL capabilities
  • provider-2/playground-v2.5 - Versatile creative model
  • provider-2/dreamshaper - Artistic dream synthesis
  • provider-2/opencole-sdxl - Open-source SDXL variant
  • provider-2/babes-by-stable-yogi-xl - Portrait specialization
  • provider-2/realism-by-stable-yogi - Photorealistic rendering
  • provider-2/realism-illustrious - Enhanced realism engine
  • provider-2/nsfw-gen - Unrestricted generation
  • provider-2/kivotos-xl - Specialized artistic style
  • provider-2/fluently-xl - Natural composition
  • provider-2/realvis-xl - Visual realism optimization
  • provider-2/flux-schnell - Rapid flux generation

• Basic Tier Models (4 models): Enhanced quality and control

  • provider-2/imagen-4.0 - Google's latest Imagen
  • provider-2/stable-diffusion-v1-5 - Industry-standard SD
  • provider-2/proteus-v0.2 - Advanced synthesis
  • provider-2/sdxl-large - Large-scale SDXL

• Pro Tier Models (8 models): Professional-grade generation

  • provider-2/dall-e-3 - OpenAI's flagship generator
  • provider-2/imagen-4.0-ultra - Maximum quality Imagen
  • provider-2/imagen-4.0-fast - Optimized Imagen
  • provider-2/flux-dev - Development flux model
  • provider-2/flux-pro-1.1 - Professional flux
  • provider-2/flux-pro-1.1-ultra - Ultra flux quality
  • provider-2/seedream-4 - Advanced dream synthesis
  • provider-2/recraft-v3 - Creative reconstruction

• Ultra Tier Exclusive (1 model): Experimental cutting-edge access

  • provider-2/nano-banana-pro - ⚠️ Limited Quota - See Critical Notice Below

IMPORTANT - Production Usage Warning: Provider 2 includes 8 models with severely limited quotas from source providers. Heavy usage may result in temporary or permanent removal.

Nano Banana Pro - Special Considerations:

• Model Classification: provider-2/nano-banana-pro is technically a chat completion model, but is currently exposed as an image generation model in Provider 2 due to source API limitations
  • When other providers release this model, it will be reclassified as a chat completion model
  • This temporary classification may change in future updates as provider support expands
• Limited Quota: Extremely restricted access from source provider
• High Removal Risk: Quota abuse will force immediate model removal from platform
• Direct Source Integration: Ensures high stability when available, but quota exhaustion is a critical concern
• Production Warning: NOT RECOMMENDED for production environments - Use for testing and experimentation only

7 Additional Limited Quota Models: Including Nano Banana Pro, there are 8 models in Provider 2 with quota restrictions. Abuse of any of these models may force platform-wide removal.

Recommendation: Avoid integration into production workflows. Use these models for testing, prototyping, and experimentation only until quota limitations are resolved.

Enhanced Image Customization: Provider 2 now supports the size parameter for select image generation models, enabling precise control over output dimensions.

• Current Availability: Limited to specific Provider 2 models
• Future Expansion: Size parameter support will be progressively rolled out across all compatible models throughout the API
• Use Case: Specify exact image dimensions for generation requirements (e.g., 1024x1024, 512x768)
• Documentation: Refer to the Image Generation API Reference for supported models and size formats

This feature addresses one of the most requested capabilities, providing developers with granular control over image output specifications.

Infrastructure Enhancements: Comprehensive stability optimizations deployed across Provider 2's infrastructure to ensure reliable performance.

• Connection Reliability: Improved request handling and error recovery mechanisms
• Response Consistency: Enhanced model routing for predictable generation times
• Error Handling: Better error messages and graceful degradation for quota-limited models
• Performance Optimization: Faster image generation processing across all models

These improvements ensure a more robust experience, particularly critical given the expanded model catalog and quota-sensitive models.

• Tier Value Enhancement: Massive value addition across all tiers—Free users gain 18 image models, Basic receives 4 specialized options, Pro unlocks 8 professional-grade models, and Ultra gains exclusive experimental access
• Creative Flexibility: 31 distinct models provide unprecedented artistic range from anime to photorealism, rapid generation to maximum quality
• Customization Control: Size parameter support marks the beginning of enhanced image customization capabilities across the platform
• Responsible Scaling: Clear communication about quota-limited models ensures sustainable platform growth while providing cutting-edge experimental access

• Quota-Limited Models: 8 models in Provider 2 have severe quota restrictions. Production use is strongly discouraged to prevent platform-wide removal.
• Nano Banana Pro Classification: Currently exposed as image generation model due to source limitations. Will be reclassified as chat completion model when broader provider support becomes available.
• Size Parameter Rollout: Currently limited to select Provider 2 models. Expansion to all compatible models across the API is planned for future updates.