Repository navigation
Refresh first-class model allowlists - #2920
antonvishal wants to merge 1 commit into
Conversation
|
|
This PR is from an external contributor and must be approved by a stagehand team member with write access before CI can run. |
|
can we break up the pr into 3? @antonvishal
|
Sure, On it. |
|
hey @miguelg719 breakup PRs are up. Closing this PR. |
…mini-3.8-flash (#2942) Mirrored from external contributor PR #2940 after approval by @miguelg719. Original author: @antonvishal Original PR: #2940 Approved source head SHA: `6503a8b9ea82b39a0a4f34d7ceaa7110c677661b` @antonvishal, please continue any follow-up discussion on this mirrored PR. When the external PR gets new commits, this same internal PR will be marked stale until the latest external commit is approved and refreshed here. ## Original description ## Summary Updates the inferred default Google Gemini model in integrations from `google/gemini-3.6-flash` to `google/gemini-3.8-flash`, and updates the Eve integration README. Part of breaking up #2920 into 3 PRs as requested in #2920 (comment) (integrations). <!-- This is an auto-generated description by cubic. --> --- ## Summary by cubic Updates the default Google model for inferred Stagehand integrations from `google/gemini-3.6-flash` to `google/gemini-3.8-flash`. - Updates the Eve integration README to document the new default. <sup>Written for commit 6503a8b. Summary will update on new commits.</sup> <a href="https://cubic.dev/pr/browserbase/stagehand/pull/2940?utm_source=github" target="_blank" rel="noopener noreferrer" data-no-image-dialog="true"><picture><source media="(prefers-color-scheme: dark)" srcset="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"><source media="(prefers-color-scheme: light)" srcset="https://www.cubic.dev/buttons/review-in-cubic-light.svg"><img alt="Review in cubic" src="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"></picture></a> <!-- End of auto-generated description by cubic. --> <!-- external-contributor-pr:owned source-pr=2940 source-sha=6503a8b9ea82b39a0a4f34d7ceaa7110c677661b claimer=miguelg719 --> Co-authored-by: Vishal Anton <vishalanton@appexert.com>
Mirrored from external contributor PR #2939 after approval by @miguelg719. Original author: @antonvishal Original PR: #2939 Approved source head SHA: `29cef2d858088d524abf510c030db139e3566204` @antonvishal, please continue any follow-up discussion on this mirrored PR. When the external PR gets new commits, this same internal PR will be marked stale until the latest external commit is approved and refreshed here. ## Original description ## Summary Refreshes model references across v4 documentation to current available model names and recommendations. Part of breaking up #2920 into 3 PRs as requested in #2920 (comment) (docs). - Updates model configuration guides and snippets to use active OpenAI, Anthropic, and Google model IDs. - Updates model recommendations in v4 documentation following Browserbase Benchmark v2. <!-- This is an auto-generated description by cubic. --> --- ## Summary by cubic Refreshes v4 docs to replace outdated model IDs with current model names and align recommendations with Browserbase Benchmark v2. Docs-only change; no SDK behavior or configuration format changes. Old IDs like `google/gemini-2.5-flash`, `anthropic/claude-sonnet-4-6`, and `openai/gpt-5.4-mini` are replaced with current names across all examples. The quick reference in `models.mdx` now recommends OpenAI models for production, speed, and cost. <sup>Written for commit 29cef2d. Summary will update on new commits.</sup> <a href="https://cubic.dev/pr/browserbase/stagehand/pull/2939?utm_source=github" target="_blank" rel="noopener noreferrer" data-no-image-dialog="true"><picture><source media="(prefers-color-scheme: dark)" srcset="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"><source media="(prefers-color-scheme: light)" srcset="https://www.cubic.dev/buttons/review-in-cubic-light.svg"><img alt="Review in cubic" src="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"></picture></a> <!-- End of auto-generated description by cubic. --> <!-- external-contributor-pr:owned source-pr=2939 source-sha=29cef2d858088d524abf510c030db139e3566204 claimer=miguelg719 --> Co-authored-by: Vishal Anton <vishalanton@appexert.com> Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
why
The allowlists still named retired OpenAI, Anthropic, and Google IDs. Docs and fixtures were pointing at the same stale names.
what changed
Protocol allowlists now keep IDs that are still on the provider APIs. Python generated models, v4 docs, and SDK/extension fixtures follow. Eval defaults were left alone so older runs stay comparable.
test plan
just generate(protocol JSON, Python models, Go client, extension zip)just fmtjust check(fmt, lint, typecheck, Python generate --check, ruff, ty, Go generate --check, gofmt, go vet)just testsdk-go,extensionassets,extensionpack, generatorSummary by cubic
Refreshes the first-class model allowlists to keep only IDs that are still reachable on the OpenAI, Anthropic, and Google APIs, and updates docs, generated SDKs, and fixtures to use current model names.
openai/gpt-6-astra,anthropic/claude-sonnet-5, andgoogle/gemini-3.8-flash.eveREADME to the new model names.google/gemini-3.6-flashtogoogle/gemini-3.8-flash.Written for commit d063084. Summary will update on new commits.