Update cerebras provider #1101

kevint-cerebras · 2025-10-01T09:00:02Z

why

The Cerebras integration was out of date when it came to models and also nonfunctional during testing.

what changed

Updated the Cerebras provider file to work with OpenAI format + added most recently released Cerebras models.

test plan

I used the form_filling_sensible file and the 2048 tests to run through all the Cerebras models in the PR

# why solves browserbase#1060 patch regression of playwright arguments being removed from agent execute response # what changed agent.execute now returns playwright arguments in its response # test plan tested locally

…ms to docs (browserbase#1065) # why reflect project id changes in docs # what changed advanced configuration comments # test plan reviewed via mintlify on localhost

# why Easier to use for Custom LLM Clients and keep users up to date with our aisdk file # what changed added export of aisdk to lib/index.ts # test plan build local stagehand, import local AISdkClient, run Azure Stagehand session

…onfigu… (browserbase#1073) …ration settings # why Updated docs to match the new fingerprint params in the Browserbase docs here: https://docs.browserbase.com/guides/stealth-customization#customization-options # what changed Update browser configuration docs to reflect the docs changes. # test plan

# why Updating docs to reflect aisdk can be imported directly # what changed The model page # test plan Reviewed page with mintlify dev locally

# why # what changed # test plan

# why Currently, we do not support stagehand agent within the api # what changed When api is enabled, stagehand agent now routes through the api # test plan Tested locally

# why Currently, using playwright screenshot command is not available when the execution environment is Stagehand. A customer has indicated they would prefer to use Playwright's native screenshot command instead of CDP when using Browserbase as CDP screenshot causes unexpected behavior for their target site. # what changed - added a StagehandScreenshotOptions type with useCDP argument added - extended page type to accept custom stagehand screeenshot options - update screenshot proxy to default useCDP to true if the env is browserbase and use playwright screenshot if false - added eval for screenshot with and without cdp # test plan - tested and confirmed functionality with eval and external example script (not committed)

…rowserbase#1057) # why We want to build a best in class agent in stagehand. Therefore, we need more eval benchmarks. # what changed - Added Web-bench evals dataset - Added a subset of OS World evals - those that can be run in a chrome browser (desktop-based tasks omitted) - added LICENSE noticed to the copied evals tasks - Added ground truth / expected result to some WebVoyager tasks using reference_answer.json from Browser Use public evals repo. Improvements to `pnpm run evals -man` to better describe how to run evals. # test plan Evals should run locally and bb for these new benchmarks.

# why Initial instructions didn't mention uv or pip prerequisites and also didn't mention venv. Fix reduces friction on first timers. # what changed - added link to install uv - added details for initializing venv - adjusted code example respectively # test plan docs change

# why - webpage structure changed, needed to update the xpath in the expected locator

… with LanguageModelV1 + LiteLLM works for python (browserbase#1086) # why 1. aisdk not yet available through npm package 2. customLLM provider only works with LanguageModelV1 3. LiteLLM compatible providers are supported in python # what changed 1. change docs to install stagehand from git repo 2. pin versions that use LanguageModelV1 # test plan local test

# why currently we pass stagehand page to agent, this results in our page management having issues when facing new tabs # what changed the stagehand object is now passed instead of stagehandPage # test plan tested locally

# why Our existing screenshot service is a dummy time-based triggered service. It also does not trigger based on any actions of the agent. # what changed Added img hash diff algo (quick check with MSE, verify with SSIM algo) to see if there was an actual UI change and only store ss in the buffer if that is so. Added ss interceptor which copies each screenshot the agent is taking to a buffer (if different enough from the previous ss) to be later used for evals. - There's also a small refactor of the agent initialization config to enable the screenshot collector service to be attached # test plan Tests pass locally --------- Co-authored-by: Miguel <[email protected]> Co-authored-by: miguel <[email protected]>

# why To help make sense of eval test cases and results # what changed Added metadata to eval runs, cleaned deprecated code # test plan

# why # what changed # test plan

# why anthropic released a new sota computer use model # what changed added claude-sonnet-4-5-20250929 as a model to the list # test plan ran evals

changeset-bot · 2025-10-01T09:00:07Z

⚠️ No Changeset found

Latest commit: 3e50149

Merging this PR will not cause a version bump for any packages. If these changes should not result in a new version, you're good to go. If these changes should result in a version bump, you need to add a changeset.

Click here to learn what changesets are, and how to add one.

Click here if you're a maintainer who wants to add a changeset to this PR

greptile-apps

Greptile Overview

Summary

This PR modernizes and fixes the Cerebras LLM provider integration by refactoring the implementation to use OpenAI-compatible format. The key changes include completely rewriting the `CerebrasClient` to use composition with the existing `OpenAIClient` rather than maintaining a custom implementation, adding 7 new Cerebras model mappings, exposing the client via exports, and completing the documentation.

The architecture change eliminates ~300 lines of duplicate code by leveraging the existing OpenAI client infrastructure while transforming Cerebras API configuration (baseURL, API keys) and model names to be compatible. The implementation strips the 'cerebras-' prefix from model names when communicating with the actual Cerebras API and properly handles logging categorization. This approach reduces maintenance burden and improves reliability by reusing battle-tested code paths from the OpenAI implementation.

Important Files Changed

Changed Files

Filename	Score	Overview
lib/llm/CerebrasClient.ts	4/5	Complete refactoring from custom implementation to OpenAI client wrapper, dramatically simplifying the codebase
lib/llm/LLMProvider.ts	2/5	Added 7 new Cerebras model mappings but model names not defined in type schema, risking validation failures
lib/index.ts	5/5	Added missing export for CerebrasClient to make it publicly available
docs/configuration/models.mdx	5/5	Added missing CEREBRAS_API_KEY documentation to complete provider setup instructions

Confidence score: 3/5

This PR improves code maintainability by eliminating duplicate implementation but has a type safety issue that could cause runtime failures
Score reflects the architectural improvement balanced against the missing type definitions for new model names in the schema validation
Pay close attention to lib/llm/LLMProvider.ts where new model names need to be added to the AvailableModelSchema type definition

Sequence Diagram

sequenceDiagram
    participant User
    participant Stagehand
    participant LLMProvider
    participant CerebrasClient
    participant OpenAIClient
    participant CerebrasAPI as "Cerebras API"

    User->>Stagehand: "new Stagehand({modelName: 'cerebras-llama-3.3-70b'})"
    Stagehand->>LLMProvider: "getClient(modelName, clientOptions)"
    LLMProvider->>LLMProvider: "Check modelToProviderMap['cerebras-llama-3.3-70b']"
    LLMProvider->>CerebrasClient: "new CerebrasClient(options)"
    CerebrasClient->>CerebrasClient: "Transform model name: remove 'cerebras-' prefix"
    CerebrasClient->>OpenAIClient: "new OpenAIClient({baseURL: 'https://api.cerebras.ai/v1'})"
    OpenAIClient-->>CerebrasClient: "OpenAI client instance"
    CerebrasClient-->>LLMProvider: "CerebrasClient instance"
    LLMProvider-->>Stagehand: "LLM client"
    Stagehand-->>User: "Stagehand instance"

    User->>Stagehand: "page.act('click button')"
    Stagehand->>CerebrasClient: "createChatCompletion(options)"
    CerebrasClient->>OpenAIClient: "createChatCompletion(options)"
    OpenAIClient->>CerebrasAPI: "POST /v1/chat/completions"
    CerebrasAPI-->>OpenAIClient: "Chat completion response"
    OpenAIClient-->>CerebrasClient: "LLM response"
    CerebrasClient-->>Stagehand: "LLM response"
    Stagehand-->>User: "Action result"

_{4 files reviewed, 1 comment}

_{Edit Code Review Agent Settings | Greptile}

lib/llm/LLMProvider.ts

kevint-cerebras · 2025-10-01T12:06:38Z

@greptileai

greptile-apps

Greptile Overview

Summary

This review covers only the changes made since the last review, not the entire PR. The developer has addressed the previous review feedback by adding the missing Cerebras model identifiers to the `AvailableModelSchema` enum in `types/model.ts`.

The change adds seven new Cerebras model identifiers: cerebras-gpt-oss-120b, cerebras-llama-4-maverick-17b-128e-instruct, cerebras-llama-4-scout-17b-16e-instruct, cerebras-qwen-3-235b-a22b-instruct-2507, cerebras-qwen-3-235b-a22b-thinking-2507, cerebras-qwen-3-32b, and cerebras-qwen-3-coder-480b. This change ensures type validation will work correctly when users specify these newer Cerebras models in their configurations.

The models are appropriately placed in the enum structure, maintaining alphabetical ordering within the Cerebras section and preserving the overall organization of the schema. This addresses the core issue identified in the previous review where the LLM provider code supported these models but the type system would reject them during validation.

Important Files Changed

Changed Files

Filename	Score	Overview
types/model.ts	5/5	Added seven new Cerebras model identifiers to AvailableModelSchema enum to enable type validation

Confidence score: 5/5

This change is safe to merge with minimal risk as it only adds new enum values without modifying existing functionality
Score reflects the straightforward nature of adding enum values and the fact that this directly addresses the previous review feedback
No files require special attention as this is a simple type schema update

_{1 file reviewed, no comments}

_{Edit Code Review Agent Settings | Greptile}

tkattkat and others added 22 commits September 10, 2025 13:40

add playwright arguments to agent (browserbase#1066)

9daa584

# why solves browserbase#1060 patch regression of playwright arguments being removed from agent execute response # what changed agent.execute now returns playwright arguments in its response # test plan tested locally

[docs] add info on not needing project id in browserbase session para…

f6f05b0

…ms to docs (browserbase#1065) # why reflect project id changes in docs # what changed advanced configuration comments # test plan reviewed via mintlify on localhost

Export aisdk (browserbase#1058)

c886544

# why Easier to use for Custom LLM Clients and keep users up to date with our aisdk file # what changed added export of aisdk to lib/index.ts # test plan build local stagehand, import local AISdkClient, run Azure Stagehand session

[docs] export aisdk (browserbase#1074)

3c39a05

# why Updating docs to reflect aisdk can be imported directly # what changed The model page # test plan Reviewed page with mintlify dev locally

Fix zod peer dependency support (browserbase#1032)

bf2d0e7

# why # what changed # test plan

add stagehand agent to api (browserbase#1077)

7f38b3a

# why Currently, we do not support stagehand agent within the api # what changed When api is enabled, stagehand agent now routes through the api # test plan Tested locally

update xpath in observe_vantechjournal (browserbase#1088)

b9c8102

# why - webpage structure changed, needed to update the xpath in the expected locator

Fix session create logs on api (browserbase#1089)

536f366

Improve failed act logs (browserbase#1090)

8ff5c5a

Eval metadata (browserbase#1092)

f89b13e

# why To help make sense of eval test cases and results # what changed Added metadata to eval runs, cleaned deprecated code # test plan

update evals cli docs (browserbase#1096)

108de3c

# why # what changed # test plan

adding support for new claude 4.5 sonnet agent model (browserbase#1099)

e0e6b30

# why anthropic released a new sota computer use model # what changed added claude-sonnet-4-5-20250929 as a model to the list # test plan ran evals

fix cerebras provider + update model list

470de96

rollback doc changes

b6759cd

add cerebras export

7f4ec6a

greptile-apps bot reviewed Oct 1, 2025

View reviewed changes

lib/llm/LLMProvider.ts Show resolved Hide resolved

update types/model.ts with new cerebras models

3e50149

greptile-apps bot reviewed Oct 1, 2025

View reviewed changes

miguelg719 force-pushed the main branch from 4994eab to bd0a799 Compare October 29, 2025 16:15

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Update cerebras provider #1101

Update cerebras provider #1101

kevint-cerebras commented Oct 1, 2025

Uh oh!

changeset-bot bot commented Oct 1, 2025 •

edited

Loading

Uh oh!

greptile-apps bot left a comment

Uh oh!

Uh oh!

kevint-cerebras commented Oct 1, 2025

Uh oh!

greptile-apps bot left a comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

9 participants

Update cerebras provider #1101

Are you sure you want to change the base?

Update cerebras provider #1101

Conversation

kevint-cerebras commented Oct 1, 2025

why

what changed

test plan

Uh oh!

changeset-bot bot commented Oct 1, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

⚠️ No Changeset found

Uh oh!

greptile-apps bot left a comment

Choose a reason for hiding this comment

Greptile Overview

Summary

Important Files Changed

Confidence score: 3/5

Sequence Diagram

Uh oh!

Uh oh!

kevint-cerebras commented Oct 1, 2025

Uh oh!

greptile-apps bot left a comment

Choose a reason for hiding this comment

Greptile Overview

Summary

Important Files Changed

Confidence score: 5/5

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

9 participants

changeset-bot bot commented Oct 1, 2025 •

edited

Loading