github lidge-jun/opencodex v2.73.0

2 hours ago

Published to npm as @bitkyc08/opencodex@2.73.0 with dist-tag latest.

New Features

  • Support Anthropic account pause and resume (#6204)
  • Add per-account Anthropic usage thresholds (#6207)
  • GPT-6.1 Sol across providers as the Sol default, and TokenLab per-model wires (#6263)

Bug Fixes

  • Redirect native tool denials through code-mode exec (#6230)
  • Avoid synthetic compaction on native/routed model switches (#6236)
  • Preserve rate-limit retry advice for Codex (carry #6225) (#6248)
  • End a passthrough stream on an upstream reset with an Anthropic error event (#6235)
  • Accept late reminders in Qwen3.8-27B translation (#6249)
  • Ship keyring native addon with desktop sidecar (#6161)
  • Strip internal summary:none marker on the wire (#6247)
  • A spent pool account must not cool the whole target (carry #6234) (#6255)
  • Accept standalone Windows wrappers in ownership probe (carry #6238) (#6258)
  • Support canonical non-streaming delivery (#6200)
  • Recover Anthropic reset cooldowns (#6203)
  • Target-platform path rules for packaged addon candidates (#6261)
  • Carry the macOS keyring signing fix to main (#6273)

Documentation

  • Record the 2.72.0 TokenLab release (#6250)

Chores

  • Open dev at 2.73.0 before releasing 2.72.0 (#6243)

Other Changes

  • Promote 2.73.0 to main (#6266)

Changelog

Full Changelog: v2.72.0...v2.73.0

  • #6161 fix(packaging): ship keyring native addon with desktop sidecar @Ingwannu
  • #6200 fix(responses): support canonical non-streaming delivery @Ingwannu
  • #6203 fix(oauth): recover Anthropic reset cooldowns @Ingwannu
  • #6204 feat(oauth): support Anthropic account pause and resume @Ingwannu
  • #6207 feat(oauth): add per-account Anthropic usage thresholds @Ingwannu
  • #6230 fix(cursor): redirect native tool denials through code-mode exec @lilinxiong
  • #6235 fix(claude): end a passthrough stream on an upstream reset with an Anthropic error event @sh940701
  • #6236 fix(catalog): avoid synthetic compaction on native/routed model switches @colthreepv
  • #6243 chore(release): open dev at 2.73.0 before releasing 2.72.0 @github-actions[bot]
  • #6247 fix(responses): strip internal summary:none marker on the wire @lidge-jun
  • #6248 fix(bridge): preserve rate-limit retry advice for Codex (carry #6225) @lidge-jun
  • #6249 fix(chat): accept late reminders in Qwen3.8-27B translation @lidge-jun
  • #6250 docs(devlog): record the 2.72.0 TokenLab release @lidge-jun
  • #6255 fix(combos): a spent pool account must not cool the whole target (carry #6234) @lidge-jun
  • #6258 fix(service): accept standalone Windows wrappers in ownership probe (carry #6238) @lidge-jun
  • #6261 fix(keyring): target-platform path rules for packaged addon candidates @lidge-jun
  • #6263 feat(models): GPT-6.1 Sol across providers as the Sol default, and TokenLab per-model wires @lidge-jun
  • #6266 release: promote 2.73.0 to main @lidge-jun
  • #6273 release: carry the macOS keyring signing fix to main @lidge-jun

Don't miss a new opencodex release

NewReleases is sending notifications on new releases.