Complete Provider Catalog Transformation & Multi-Provider Expansion

Strategic Repositioning: Complete catalog replacement focusing on cutting-edge open-source and emerging AI lab models, retiring all previous OpenAI GPT and Google Gemini offerings for specialized model diversity.

30 New Models Added:

• Free Tier Models: Accessible flagship models for all users
  • Meta Llama Family:
    • provider-1/meta-llama-3.3-70b-instruct - 70B flagship
    • provider-1/meta-llama-3.1-8b-instruct - Efficient 8B variant
    • provider-1/llama-3.1-8b-instruct - Alternative 8B deployment
    • provider-1/llama-4-scout-17b-16e-instruct - Next-gen 17B model
    • provider-1/llama-4-maverick-17b-128e-instruct - 17B Mixture-of-Experts
  • Qwen3 Series:
    • provider-1/qwen3-32b - 32B flagship
    • provider-1/qwen3-30b-a3b - 30B variant
  • Other Models:
    • provider-1/exaone-4.0.1-32b - LG AI Research 32B
    • provider-1/magistral-small-2506 - Mistral AI compact variant
    • provider-1/devstral-small-2505 - Mistral AI developer-focused
    • provider-1/gemma-3-27b-it - Google Gemma 27B instruction-tuned
    • provider-1/ax-3.1 - AX 3.1 model
    • provider-1/allam-7b-instruct-preview - 7B preview model
    • provider-1/deepseek-r1-0528 - DeepSeek R1 reasoning
    • provider-1/deepseek-r1-distill-llama-70b - DeepSeek R1 70B distilled
• Basic Tier Models: Enhanced capabilities and advanced models
  • Qwen3 Advanced:
    • provider-1/qwen3-235b-a22b-instruct-2507 - 235B flagship
    • provider-1/qwen3-235b-a22b-thinking-2507 - 235B thinking variant
  • DeepSeek V3 Series:
    • provider-1/deepseek-v3-0324 - V3 0324 checkpoint
    • provider-1/deepseek-v3.1 - V3.1 flagship
    • provider-1/deepseek-v3.1-terminus - Hybrid reasoning variant
  • Specialized Models:
    • provider-1/llama-3.3-swallow-70b-instruct-v0.4 - Specialized 70B Llama variant
    • provider-1/mistral-small-3.1-24b-instruct-2503 - Mistral AI 24B flagship
    • provider-1/ax-4.0 - AX 4.0 flagship
    • provider-1/hyperclovax-seed-think-14b - HyperCLOVA X thinking model
    • provider-1/gpt-oss-120b - OpenAI 120B open-source
  • Audio Transcription:
    • provider-1/whisper-large-v3 - OpenAI Whisper for speech-to-text
• Pro Tier Models: Premium flagship
  • provider-1/glm-4.6 - Zhipu AI GLM 4.6

13 Models Removed:

• Complete OpenAI Retirement: All GPT models removed from Provider 1 catalog
  • provider-1/gpt-5-pro, provider-1/gpt-5-nano, provider-1/gpt-5-mini, provider-1/gpt-5
  • provider-1/gpt-4.1, provider-1/gpt-4.1-nano, provider-1/gpt-4.1-mini
  • provider-1/gpt-4o, provider-1/gpt-4o-mini, provider-1/o4-mini
• Google Gemini Retirement: provider-1/gemini-2.5-flash
• Mistral AI Retirement: provider-1/mistral-large-latest, provider-1/codestral-latest, provider-1/magistral-medium-latest

8 New Models Added:

• Free Tier Models: Accessible flagship models
  • provider-3/glm-4.5-free - Zhipu AI's GLM 4.5
• Basic Tier Models: Enhanced capabilities
  • provider-3/qwen-3-next-80b-a3b-thinking - Qwen 80B thinking model
  • provider-3/glm-4.6-free - Zhipu AI's GLM 4.6
  • provider-3/gemini-2.5-flash-preview-09-2025 - Google Gemini 2.5 Flash preview
• Pro Tier Models: Advanced reasoning and image editing
  • provider-3/kimi-k2-thinking - Moonshot AI's K2 reasoning model
  • provider-3/flux-krea-dev - Black Forest Labs image editing
• Ultra Tier Exclusive: Premium flagship models
  • provider-3/claude-4.5-haiku - Anthropic's fast, efficient Haiku
  • provider-3/grok-3-mini - xAI's compact Grok variant

3 Models Removed:

• gpt-4o-mini-search-preview - Search-enabled GPT-4o mini variant
• gpt-4o-search-preview - Search-enabled GPT-4o variant
• text-moderation-latest - OpenAI text moderation model

Model Upgrade:

• Free Tier Models: Next-generation image generation
  • provider-4/imagen-3.5 - Google's enhanced Imagen 3.5

Model Removed:

• Imagen 3 Retirement: provider-4/imagen-3 - Superseded by imagen-3.5

• Free Tier Models: Accessible image generation and chat capabilities
  • Image Generation:
    • provider-5/imagen-4-fast - Google's fast Imagen 4 variant
    • provider-5/nanobanana - Efficient image generation
  • Chat Models:
    • provider-5/polaris-alpha - Free flagship model
• Basic Tier Models: Enhanced generation and chat capabilities
  • Image Generation:
    • provider-5/flux-krea - Black Forest Labs creative generation
    • provider-5/imagen-4-preview - Google's Imagen 4 preview with enhanced quality
  • Chat Models:
    • provider-5/minimax-m2 - MiniMax M2 flagship
    • provider-5/kimi-linear-48b-a3b-instruct - Moonshot AI's 48B model
• Pro Tier Models: Advanced reasoning capabilities
  • provider-5/kimi-k2-thinking - Moonshot AI's K2 reasoning model with function calling

Platform Architecture:

• Provider 1 Strategic Refocus: Transformed from general-purpose to specialized open-source and emerging AI lab hub, removing all proprietary closed-source models (OpenAI GPT, Google Gemini) for focused catalog of cutting-edge research models
• Multi-Provider Optimization: Enhanced load balancing and intelligent routing across providers for improved reliability and reduced latency
• Model Diversity: 30+ new models from Meta, Qwen, DeepSeek, and emerging AI labs across Provider 1, with 8 flagship additions to Provider 3 and 8 new models in Provider 5
• Tier-Based Access: Optimized model distribution across Free, Basic, Pro, and Ultra tiers for sustainable resource allocation while maintaining accessibility

• Provider 7 Streaming Fix: Resolved streaming issues with Claude models in Provider 7, ensuring smooth real-time response delivery
• Provider 5 Image Models: Fixed image generation models in Provider 5 for improved reliability and performance
• Website Enhancements: Various small modifications and bug fixes across the platform for better user experience and stability

• Breaking Changes: Provider 1 users relying on GPT or Gemini models must migrate to Provider 5 for OpenAI models or Provider 3 for Google models. All affected model IDs have been completely removed from Provider 1.
• Imagen Upgrade: Users of provider-4/imagen-3 should update to provider-4/imagen-3.5 for enhanced image generation quality and capabilities.
• Audio Capabilities: Provider 1 now includes Whisper Large V3 for high-quality speech-to-text transcription, expanding multimodal capabilities beyond text generation.