v0.8.3 v0.1.184
cheaper-inference/OmniRouteAug 30, 2026 → Aug 31, 20261 day between releases
What's New
New release introduces a comprehensive AI API gateway overhaul with Codex routing, public group access control, and multi-provider support.
New Features
- Codex routing model directory with precise account model aliases
- Native statistics for compaction requests and pre-inference intensity mapping
- Admin-controlled access restrictions for public groups
- Support for Zhipu Team GLM Coding Plan usage query
- Ollama Cloud usage window mounting for domestic platforms
Improvements
- Optimized WebSocket, sticky sessions, and large request forwarding
- Optimized Codex model directory cache, capability sync, and discovery
- Optimized DeepSeek peak/off-peak billing prices
- Optimized account expiration time local timezone parsing
- Optimized upstream rate limit detection to prevent page refreshes
Bug Fixes
- Fixed Anthropic to Responses streaming thinking content block order and index
- Fixed Anthropic/Bedrock transport layer error fault transfer trigger
- Fixed Grok Responses tool output and invalid tool joint type handling
- Fixed OpenAI streaming failure, quota reset, and client disconnect error judgment
- Fixed OAuth re-authorization and coupon retention issues
Upgrade Notes
The upgrade introduces a completely new feature set including Codex routing and public group access control. Ensure configuration supports the new Zhipu and Ollama providers and verify WebSocket settings. Review the new admin-controlled access restrictions for public groups.
The Two Releases
v0.8.3
Aug 30, 2026· 2 months ago
A maintenance release focusing on CI alignment, fixing Action Ring selection issues, and repairing i18n gate and agent guidance.
Full release notesSub2API 0.1.184
Aug 31, 2026· 2 months ago
This release focuses on enhancing the Codex routing model directory, adding native usage statistics for compaction requests, and improving gateway stability and access control for public groups.
Full release notes