fix(thinking): align budget effort mapping across translators

feat(iflow): add thinking support for iFlow models

refactor(thinking): export thinking helpers

fix(thinking): drop unsupported none effort

Fixed: #534

Merge pull request #449 from sususu98/fix/gemini-cli-429-retry-delay-parsing

fix(executor): add `allowCompat` support for reasoning effort normalization

Merge branch 'dev' into think

Merge pull request #501 from huynguyen03dev/fix/openai-compat-model-alias-resolution

fix(executor): improve model compatibility handling for OpenAI-compatibility

fix(thinking): map budgets to effort levels

fix(openai-compat): prevent model alias from being overwritten by ResolveOriginalModel

fix(codex): raise default reasoning effort to medium

refactor(executor): relocate gemini token counters

refactor(executor): clarify executor comments and oauth names

refactor(executor): clarify providers and streams

Merge pull request #481 from sususu98/fix/increase-buffer-size

refactor(thinking): use parentheses for metadata suffix

refactor(thinking): use bracket tags for thinking meta

fix(runtime): unify claude thinking config resolution

docs(runtime): document reasoning effort precedence

fix(runtime): validate thinking config in iflow and qwen

fix(runtime): unify reasoning effort metadata overrides

fix(runtime): validate reasoning effort levels

feat(runtime): add thinking config normalization

fix(gemini-cli): enhance 429 retry delay parsing

feat(util): implement dynamic thinking suffix normalization and refactor budget resolution logic

fix(executor): increase buffer size for stream scanners to 50MB across multiple executors

fix(logging): update response aggregation logic to include all attempts

fix(antigravity): remove references to `autopush` endpoint and update fallback logic

fix(executor): centralize default thinking config

feat(executor): enforce minimum thinking budget for antigravity models

feat(executor): normalize thinking budget across all Gemini executors

Merge pull request #470 from router-for-me/agry

Fixed: #463

feat(antigravity): support canonical names for antigravity models

feat(aistudio): normalize thinking budget in request translation

fix(antigravity): remove `exclusiveMaximum` from JSON during key deletion

refactor(logging): remove unused variable in `ensureAttempt` and redundant function call

feat(antigravity): enforce thinking budget limits for Claude models

feat(registry): add explicit thinking support config for antigravity models

Fixed: #424

fix(antigravity): restore production base URL in the executor

Fixed: #421

**refactor(antigravity): handle `anyOf` property, remove `exclusiveMinimum`, and comment unused prod URL**

**feat(util): add DeleteKey function and update antigravity executor for Claude model compatibility**

Fixed: #414

**feat(antigravity): add support for Claude-Opus-4-5-Thinking model**

refactor(antigravity): uncomment prod environment URL in fallback chain

**refactor(cliproxy, config): remove vertex-compat flow, streamline Vertex API key handling**