A4F
v19.5.0•

Provider 6 & 8 Launch, Flux 2 Series & Major Infrastructure Updates

Sreejan
SreejanAuthor

This update introduces two new providers to the platform with an extensive catalog of models, while addressing ongoing quota challenges across multiple providers. Provider 6 brings Groq-powered high-performance models, and Provider 8 debuts with a comprehensive suite of 36 chat and image generation models. We have also implemented strategic adjustments to Provider 2 and Provider 3 in response to quota limitations.

⚠️ Important Service Notices

4

Critical Updates Regarding Model Availability:

  • GPT Model Availability: We regret to inform our users that GPT models have been experiencing significant quota exhaustion across multiple providers. Our team is actively working to secure new provider partnerships to restore GPT model access. We appreciate your patience during this transition period and will issue an update as soon as new providers are onboarded.

  • Gemini Model Status: Due to strict API restrictions imposed by Google, Gemini model availability remains limited and should be considered experimental. Quota constraints continue to affect service reliability, and users are advised to have fallback options in place when utilizing Gemini models.

  • Claude Model Stability: We are pleased to report that Claude models remain stable and fully operational as usual. No service disruptions have been reported. However, please note that the newly added Claude models in Provider 8 are from an experimental provider and may not exhibit the same stability as Provider 7.

  • Free Tier Compensation: To mitigate any inconvenience caused by the ongoing quota challenges, we have ensured extensive Free tier access across Provider 6 and Provider 8. We encourage all users to explore these new providers for continued access to high-quality AI models.

🔧 Roo Code Compatibility Notice

1

XML Function Deprecation in Roo Code:

Roo Code has deprecated the XML function calling format and implemented native function calling exclusively. Users experiencing compatibility issues or unexpected behavior are advised to downgrade to Roo Code version 3.16.x until we release a compatibility fix.

Our development team is actively working on addressing this change, and a fix will be deployed in an upcoming update. We sincerely apologize for any inconvenience this may cause and appreciate your understanding.

Provider 6 - Groq-Powered Model Launch

11

High-Performance Groq Integration: Provider 6 debuts with 11 models optimized for speed and efficiency through Groq's specialized inference infrastructure.

Free Tier Models (8 models):

Basic Tier and Above (3 models):

Provider 8 - Comprehensive Model Suite Launch

36

New Multi-Purpose Provider: Provider 8 introduces 36 models spanning chat, reasoning, and image generation, offering extensive Free tier access.

Chat & Reasoning Models (27 models)

Free Tier Models (20 models):

Basic Tier and Above (5 models):

Pro Tier and Above (2 models):

Image Generation Models (9 models)

All Tiers - Free, Basic, Pro, Ultra:

Provider 2 - Flux 2 Image Generation Suite

4

Flux 2 Series Retained: Despite quota exhaustion affecting GPT and Gemini models, the Flux 2 image generation suite remains fully operational and serves as the cornerstone of Provider 2.

⚠️ Models Removed Due to Quota Exhaustion:

  • All GPT models previously available on Provider 2 have been removed due to complete quota exhaustion
  • All Gemini models previously available on Provider 2 have been removed due to complete quota exhaustion

Active Flux 2 Models:

  • provider-2/flux-2-dev - Pro & Ultra Tiers

    • Development-grade model with maximum configurability
    • Generation Config: Steps: 28, Guidance Scale: 3.5
    • Price: $0.025 per image
    • Ideal for iterative creative workflows and prototyping

  • provider-2/flux-2-pro - Pro & Ultra Tiers

    • Professional-grade model with integrated safety controls
    • Generation Config: Safety Tolerance: 2
    • Price: $0.05 per image
    • Optimized for production-ready outputs with content moderation

  • provider-2/flux-2-flex - Pro & Ultra Tiers

    • Most configurable model supporting all generation parameters
    • Generation Config: Steps: 28, Guidance Scale: 3.5, Safety Tolerance: 2
    • Price: $0.035 per image
    • Maximum flexibility for advanced users requiring full parameter control

  • provider-2/flux-2-max - Pro & Ultra Tiers

    • Maximum quality model for premium image generation
    • Generation Config: Safety Tolerance: 2
    • Price: $0.06 per image
    • Highest fidelity outputs for professional creative applications

The Flux 2 series represents the most important addition to Provider 2 and helps sustain the provider's value proposition for image generation workflows.

Provider 3 - Service Status Update

1

Gemini Models Temporarily Unavailable:

We regret to inform our users that Gemini models on Provider 3 have become temporarily unavailable due to provider-side resource constraints. However, GPT models on Provider 3 remain fully operational.

  • Gemini Models: Currently unavailable. The provider has indicated that an update is forthcoming, and we anticipate Gemini services to be restored in a future release.
  • GPT Models: Fully operational with no reported issues.

We will issue an update as soon as Provider 3 restores Gemini model availability.

Backend Optimizations

1

Dependency Cleanup: Removed unused aioredis dependency from the main application module, streamlining the codebase and reducing unnecessary imports. This optimization improves application startup time and reduces memory footprint without affecting functionality.

Platform Impact

5
  • New Provider Expansion: Provider 6 and Provider 8 significantly expand the platform's model catalog with 47 new models across chat, reasoning, and image generation
  • Free Tier Enhancement: Extensive Free tier access across new providers ensures continued AI access despite quota challenges on legacy providers
  • Image Generation Strength: Flux 2 series on Provider 2 combined with 9 image models on Provider 8 provides comprehensive image generation capabilities
  • Groq Performance: Provider 6's Groq integration delivers optimized inference speeds for latency-sensitive applications
  • Multi-Provider Redundancy: New providers offer alternative access paths for models affected by quota exhaustion on other providers

Important Notes

6
  • GPT Model Restoration: Our team is actively seeking new provider partnerships to restore GPT model availability. Updates will be communicated as soon as new sources are secured.
  • Gemini Experimental Status: All Gemini models should be considered experimental due to Google's strict API quota restrictions. Please plan accordingly with fallback options.
  • Claude Provider 8 Notice: Claude Sonnet 4.5 on Provider 8 is from an experimental provider. For maximum stability, we recommend using Provider 7 for Claude models.
  • Free Tier Compensation: Provider 6 and Provider 8 offer extensive Free tier access to compensate for quota-related disruptions. We encourage users to explore these new options.
  • Roo Code Users: If experiencing issues with the new native function calling format, please downgrade to Roo Code version 3.16.x. A compatibility fix is in development.
  • Flux 2 Pricing: Flux 2 models use per-image pricing ranging from $0.025 to $0.06 based on model tier and output quality requirements.