v0.9.11 v0.1.180
cheaper-inference/OmniRouteAug 23, 2026 → Aug 24, 20261 day between releases
What's New
New release introduces OpenAI reset cards and OAuth plugin, diverging from previous Codewhale CLI tools.
Breaking Changes
- Legacy npm package `deepseek-tui` is deprecated
New Features
- OpenAI reset card auto-usage based on thresholds
- Fast mode service_tier billing for OpenAI
- Configurable limit for upstream model list responses
- Model square with context tier pricing
- Experimental OAuth outbound transmission plugin
Improvements
- Unified token billing path and standardized Anthropic usage
- Updated Grok 4.6 model directory and Realtime protocol
- OAuth account auto-sync for model discovery
- DeepSeek client tool migration to native Responses
- Improved deployment configuration and OpenAI force_http handling
Bug Fixes
- Fixed illegal tool parameters and file content loss in OpenAI bridge
- Resolved gateway pool mode errors and capacity overload
- Fixed Grok 429, compaction 422, and handshake compatibility issues
- Fixed vendor protocol routing and Anthropic billing anomalies
- Fixed Google One model constraints and XSS bypass in security dependencies
Upgrade Notes
Migrate from the deprecated `deepseek-tui` package to the new `codewhale` or `sub2api` entrypoints. Note the shift in Docker image from `hmbown/codewhale` to `weishaw/sub2api`.
The Two Releases
v0.9.11
Aug 23, 2026· 2 months ago
This release introduces the Codewhale rebrand, adds vision-capable DeepSeek V4 Flash models, and introduces a Codewhale-versus-Pi evaluation harness. It also improves terminal UX and fixes workflow visibility issues.
Full release notesSub2API 0.1.180
Aug 24, 2026· 2 months ago
Introduces OpenAI reset card auto-usage, Fast mode service_tier billing, configurable model list size limits, and improvements to the model square and vendor compatibility.
Full release notes