v18.2.0•
Provider 5 Catalog Optimization & GPT-5.1 Expansion
Priya ShekhawatAuthor
This strategic update optimizes Provider 5's catalog by removing 74 redundant models while introducing OpenAI's next-generation GPT-5.1 family, advanced image generation capabilities, comprehensive embedding models, and text-to-speech enhancements. Additionally, we've resolved critical reasoning support issues for Claude Sonnet 4.5.
New Features
11
- GPT-5.1 Family - Ultra Tier Exclusive: Next-generation adaptive reasoning models with multimodal support
- Core Chat Models:
provider-5/gpt-5.1- Advanced GPT-5.1 with 400K context window, 128K max outputprovider-5/gpt-5.1-chat-latest- Latest chat-optimized variant with instant reasoning mode
- Dated Variant:
provider-5/gpt-5.1-2025-11-13- November 2025 snapshot
- Responses API Specialists:
provider-5/gpt-5.1-codex- Code generation specialist optimized for/v1/responsesendpointprovider-5/gpt-5.1-codex-mini- Efficient coding variant optimized for/v1/responsesendpoint
- Features: Reasoning, function calling, vision capabilities
- Performance: 2.0-6.5s latency, 55-75 tokens/sec throughput, 97.85-98.65% uptime
- Core Chat Models:
- Image Generation Models - Pro & Ultra Tiers: Advanced realistic image generation by OpenAI
provider-5/gpt-image-1- Premium quality with multiple quality levels (Low/Medium/High: $0.011/$0.042/$0.167 per image)provider-5/gpt-image-1-mini- Cost-effective variant (Low/Medium/High: $0.005/$0.011/$0.036 per image)- Features: Up to 4 images per request, 3-12s latency, 98.25-98.75% uptime
- Text Embedding Models: High-performance semantic embeddings for RAG applications
provider-5/text-embedding-3-small- 1536 dimensions, $0.02 per 1M tokens (Basic, Pro, Ultra)provider-5/text-embedding-3-large- 3072 dimensions, $0.13 per 1M tokens (Pro, Ultra only)- Features: 8191 token context window, 280-350 tokens/sec throughput, 99.45-99.65% uptime
- Text-to-Speech Model - Pro & Ultra Tiers: High-quality voice synthesis
provider-5/gpt-4o-mini-tts- Cost-effective TTS at $0.015 per 1K characters- Features: Multiple voice options, 0.4-1.5s latency, 185 chars/sec throughput, 99.35% uptime
Models Removed from Provider 5
74
- xAI Grok Models (7 models): All Grok variants removed from Provider 5
- Anthropic Claude Models (6 models): All Claude variants removed from Provider 5
- Qwen Models (16 models): Complete Qwen3 series removal from Provider 5
qwen3-235b-a22b-2507,qwen3-coder,qwen3-vl-235b-a22b-instruct,qwen3-next-80b-a3b-instruct,qwen3-30b-a3b-instruct-2507,qwen3-32b,qwen3-235b-a22b-thinking-2507,qwen3-30b-a3b,qwen-2.5-72b-instruct,qwen-2.5-7b-instruct,qwen2.5-vl-32b-instruct,qwen3-max,qwen3-14b,qwen3-235b-a22b,qwen3-coder-exacto,qwen3-coder-30b-a3b-instruct
- DeepSeek Models (9 models): Complete DeepSeek lineup removed from Provider 5
- Perplexity Sonar Models (6 models): All Sonar web search models removed from Provider 5
- Amazon Nova Models (4 models): Complete Nova series removed from Provider 5
- Google Gemini Models (8 models): Gemini 2.5 and 2.0 series removed from Provider 5
- OpenAI Search Models (2 models): Web search preview models removed from Provider 5
- Other Provider Models (16 models): Miscellaneous third-party models removed
- OpenRouter:
polaris-alpha - MiniMax:
minimax-m2 - Moonshot AI:
kimi-linear-48b-a3b-instruct,kimi-k2-thinking - Additional models from various providers
- OpenRouter:
- Strategic Rationale: Provider 5 now focuses exclusively on OpenAI's native offerings with direct API integration for optimal performance, lowest latency, and highest reliability
Bug Fixes
1
- Provider 7 Reasoning Support Fixed: Resolved critical issue with
provider-7/claude-sonnet-4-5-20250929where some users were unable to enable reasoning support. The model now correctly handles reasoning mode activation for all subscription tiers.
Migration Notes
2
- Model Availability: All removed models remain accessible through Provider 1 and Provider 3 with full feature parity and identical capabilities
- Grok Users: xAI Grok models continue to be available through Provider 3