The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
litellm	Building	Preview, Comment	Mar 18, 2026 0:38am

Merging this PR will not alter performance

✅ 16 untouched benchmarks

_{Comparing Sameerlite:litellm_vertex_paygo_tutorial (ea80a19) with main (cec3e9e)}

Greptile Summary

This PR adds a new documentation tutorial (vertex_ai_pay_go.md) covering Vertex AI Priority PayGo and Standard PayGo vs Provisioned Throughput, an accompanying SVG cost-tracking flow diagram, and a sidebar entry under "Spend Tracking". The content is well-structured and the code examples are technically accurate, but there are a couple of accuracy/completeness issues worth addressing before merge.

• Provider mismatch in supported models list: The info callout lists gemini/gemini-2.5-pro alongside vertex_ai/ models as supporting Priority PayGo. The gemini/ prefix routes through Google AI Studio, not Vertex AI, so the X-Vertex-AI-LLM-Shared-Request-Type: priority header won't have the intended effect — this should be vertex_ai/gemini-2.5-pro.
• Incomplete cost table: The trafficType → service_tier mapping table only documents input_cost_per_token_* keys; output_cost_per_token_priority and output_cost_per_token_flex keys are also present in the model pricing JSON and should be listed for completeness.
• SVG missing trailing newline: Minor — the vertex_cost_tracking_flow.svg file has no newline at end of file.

Confidence Score: 4/5

• Documentation-only PR; safe to merge after fixing the provider prefix inaccuracy in the supported models callout.
• No code changes are involved — all three files are documentation and static assets. The main concern is a factual inaccuracy (listing a gemini/ provider model in a Vertex AI PayGo context) that could mislead users, but it does not affect runtime behavior. The fix is a one-line change.
• docs/my-website/docs/tutorials/vertex_ai_pay_go.md — line 12 (provider prefix) and lines 87-91 (incomplete cost table)

Flowchart

%%{init: {'theme': 'neutral'}}%%
flowchart TD
    A["HTTP Request\nX-Vertex-AI-LLM-Shared-Request-Type: priority"] -->|Vertex AI| B["Vertex Response\nusageMetadata.trafficType = ON_DEMAND_PRIORITY"]
    B --> C["LiteLLM stores it\n_hidden_params.provider_specific_fields.traffic_type"]
    C --> D["completion_cost()\nMaps traffic_type → service_tier = 'priority'"]
    D --> E["Pricing lookup\ninput/output_cost_per_token_priority"]

    style A fill:#0c447c,stroke:#85b7eb,color:#b5d4f4
    style B fill:#085041,stroke:#5dcaa5,color:#9fe1cb
    style C fill:#3c3489,stroke:#afa9ec,color:#cecbf6
    style D fill:#633806,stroke:#ef9f27,color:#fac775
    style E fill:#712b13,stroke:#f0997b,color:#f5c4b3

_{Last reviewed commit: "Fix greptile review"}