docs(oci): expand provider page with proxy, tools, vision, reasoning, env vars#197
Merged
Merged
Conversation
… env vars Bring the OCI provider docs up to parity with the Bedrock page: - Environment variables for all credentials (OCI_REGION, OCI_USER, OCI_FINGERPRINT, OCI_TENANCY, OCI_COMPARTMENT_ID, OCI_KEY, OCI_KEY_FILE) - LiteLLM Proxy Usage section: config.yaml example with both Grok and Cohere entries, start command, Curl + OpenAI client smoke tests - Function Calling / Tool Calling: OpenAI-compatible tools example for both SDK and proxy modes, with a note that Cohere and Generic vendors are adapted internally - Vision / Multimodal: image_url example plus the full list of vision-capable models - Reasoning / Thinking: reasoning_effort (low/medium/high/disable) and reasoning_tokens surfaced on usage; documents that the param is silently ignored for Cohere models - Optional Parameters table extended with an Environment Variable column and a reasoning_effort row Reconciled the Supported Models list against OCI's on-demand retirement page: - Removed retired meta.llama-3.1-405b-instruct and meta.llama-3.1-70b-instruct - Added xai.grok-4.3, openai.gpt-oss-120b/20b, and cohere.embed-multilingual-image-v3.0 - Flagged retirement dates on Llama 3.2-90b-vision, all Grok 3/4/4.x, Cohere R+/R 08-2024, and all embed v3.0 models - Switched the Vision example from Llama 3.2-90b (retires 2026-09-30) to Llama 4 Maverick - Added an info callout linking to OCI's retirement page so readers can verify dates themselves
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
Other LiteLLM provider pages (bedrock, openai, anthropic) don't track retirement dates inline — they list active models and let the provider's own lifecycle page own the dates. Drop the per-model *(retires ...)* annotations to match. The single info-callout link to OCI's retirement page stays so readers can verify the schedule.
Collaborator
|
bugbot run |
![cursor[bot]](https://avatars.githubusercontent.com/in/1210556?s=60&v=4){kind=small}
There was a problem hiding this comment.
✅ Bugbot reviewed your changes and found no new issues!
Comment @cursor review or bugbot run to trigger another review on this PR
Reviewed by Cursor Bugbot for commit 127c7b5. Configure here.
mateo-berri
approved these changes
May 23, 2026
Collaborator
mateo-berri
left a comment
mateo-berri
left a comment
There was a problem hiding this comment.
Thanks for the contributions! LGTM
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Companion docs for the OCI Generative AI integration in BerriAI/litellm#25177.
The existing
docs/providers/oci.mdcovered manual credentials, SDK signer, streaming, dedicated endpoints, and embeddings — but was missing several sections that the Bedrock page (the closest "first-class provider" reference) already had. This PR brings OCI up to parity and reconciles the model list against OCI's authoritative retirement page.What's new
OCI_REGION,OCI_USER,OCI_FINGERPRINT,OCI_TENANCY,OCI_COMPARTMENT_ID,OCI_KEY,OCI_KEY_FILE) — already wired up inlitellm/llms/oci/common_utils.py.config.yamlexample with both Grok and Cohere entries, start command, and Curl + OpenAI client smoke tests (mirrors the Bedrock section).image_urlexample plus the full list of vision-capable models.reasoning_effort(low/medium/high/disable) andreasoning_tokenssurfaced onusage; documents that the parameter is silently ignored for Cohere models (they're not reasoning models).Environment Variablecolumn and areasoning_effortrow.Model catalog corrections
Reconciled against OCI's on-demand retirement page:
meta.llama-3.1-405b-instructandmeta.llama-3.1-70b-instruct.xai.grok-4.3(OCI's recommended replacement for the entire Grok 3/4 family),openai.gpt-oss-120b/openai.gpt-oss-20b, andcohere.embed-multilingual-image-v3.0.embed-*-v3.0models (2026-09-30).:::infocallout linking to OCI's retirement page so readers can verify dates themselves as the catalog evolves.cc @mateoo
Note
Low Risk
Documentation-only update that expands and corrects OCI provider guidance; low risk aside from potential user confusion if any model names/params are inaccurate.
Overview
Expands
docs/providers/oci.mdwith new OCI usage guidance: environment-variable based auth, LiteLLM Proxyconfig.yaml+ curl/OpenAI client examples, and OpenAI-compatible tool/function calling examples for both SDK and proxy.Updates the supported model catalog (adds new Grok/OpenAI OSS/embedding entries and marks multimodal-capable models), and adds vision/multimodal and reasoning (
reasoning_effort) sections including how reasoning tokens surface inusage. The optional-parameters table is revised to include an Environment Variable column and areasoning_effortrow, plus a link to OCI’s official model retirement page.Reviewed by Cursor Bugbot for commit 127c7b5. Bugbot is set up for automated code reviews on this repo. Configure here.