Skip to content

docs: add v1.82.3 release notes#23820

Merged
7 commits merged intoBerriAI:mainfrom
joereyna:release-notes/v1.82.3-v2
Mar 18, 2026
Merged

docs: add v1.82.3 release notes#23820
7 commits merged intoBerriAI:mainfrom
joereyna:release-notes/v1.82.3-v2

Conversation

@joereyna
Copy link
Copy Markdown
Contributor

@joereyna joereyna commented Mar 17, 2026

Summary

Re-submission of #23816 (which was reverted) with the following fixes:

Changes

  • docs/my-website/release_notes/v1.82.3.md — full release notes (116 new models, 132 removed, 5 new providers)

….json

- Fix provider count header: 4 -> 5 new providers
- Fix nebius/zai: add bridged endpoint support (messages, responses, a2a, interactions)
- Add missing sagemaker_nova to provider_endpoints_support.json
@vercel
Copy link
Copy Markdown

vercel bot commented Mar 17, 2026

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

Project Deployment Actions Updated (UTC)
litellm Ready Ready Preview, Comment Mar 18, 2026 5:19am

Request Review

@codspeed-hq
Copy link
Copy Markdown
Contributor

codspeed-hq bot commented Mar 17, 2026

Merging this PR will not alter performance

✅ 16 untouched benchmarks


Comparing joereyna:release-notes/v1.82.3-v2 (8a4ef0b) with main (cfeafbe)

Open in CodSpeed

@greptile-apps
Copy link
Copy Markdown
Contributor

greptile-apps bot commented Mar 17, 2026

Greptile Summary

This PR re-submits the v1.82.3 release notes (reverted PR #23816) with four targeted corrections: provider count fixed to 5, Nebius AI PR link corrected to #22614, ZAI docs link updated to the dedicated /docs/providers/zai page (which confirmed exists), and the duplicate PR link for SageMaker Nova resolved by keeping #21542 distinct from Nebius AI's #22614. All four claimed fixes are verified in the diff.

Key observations:

  • All four stated fixes are correctly applied and verified against the repository (dedicated zai.md and nebius.md docs pages confirmed to exist)
  • The release notes cover a large surface area: 116 new models, 5 new providers, security patches, and performance fixes
  • One minor style inconsistency remains: the Key Highlights bullet describes gpt-5.4 with "1M context" while the models table lists "1.05M" for the same model

Confidence Score: 4/5

  • This is a documentation-only PR; no runtime code is changed, making it safe to merge with minor cleanup.
  • All four fixes from the PR description are correctly applied and independently verified. The only remaining issue is a minor intra-document inconsistency (1M vs 1.05M context window) that has no functional impact. No code, schema, or configuration files are touched.
  • No files require special attention — this is a docs-only change.

Important Files Changed

Filename Overview
docs/my-website/release_notes/v1.82.3.md New release notes file documenting v1.82.3 changes — all four fixes from the reverted PR #23816 are correctly applied (5 providers, correct Nebius AI PR link, dedicated ZAI doc page, distinct SageMaker Nova PR link); one minor contextual inconsistency found between Key Highlights and model table for gpt-5.4 context window.

Flowchart

%%{init: {'theme': 'neutral'}}%%
flowchart TD
    A["v1.82.3 Release Notes"] --> B["New Providers (5)"]
    A --> C["New Models (116)"]
    A --> D["Removed Models (132)"]
    A --> E["Features & Bug Fixes"]
    A --> F["Security Updates"]

    B --> B1["Nebius AI — nebius/"]
    B --> B2["ZAI — zai/"]
    B --> B3["Black Forest Labs — black_forest_labs/"]
    B --> B4["Serper — serper/"]
    B --> B5["SageMaker Nova — sagemaker_nova/"]

    C --> C1["OpenAI gpt-5.4 / gpt-5.4-pro"]
    C --> C2["Gemini 3.x models"]
    C --> C3["FLUX Kontext image editing"]
    C --> C4["Mistral Magistral reasoning"]
    C --> C5["Dashscope Qwen3 VL"]
    C --> C6["30 Nebius AI models"]

    F --> F1["Secret redaction in proxy logs"]
    F --> F2["PyJWT bump to ^2.12.0"]
    F --> F3["tar + tornado CVE patches"]
Loading

Comments Outside Diff (1)

  1. docs/my-website/release_notes/v1.82.3.md, line 46 (link)

    Context window mismatch: "1M" vs "1.05M"

    The Key Highlights bullet describes gpt-5.4 as having "1M context", but the models table below lists its context window as 1.05M. These values are inconsistent within the same document. Consider either rounding the table entry to 1M or updating the Key Highlights to read "1.05M context" so readers get a consistent figure.

Last reviewed commit: "revert: restore full..."

Copy link
Copy Markdown
Contributor

@yuneng-jiang yuneng-jiang left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Please confirm if the provider_endpoints_support.json changes are intentional

@joereyna
Copy link
Copy Markdown
Contributor Author

Please confirm if the provider_endpoints_support.json changes are intentional

Yes, intentional since sagemaker provider was added.

@joereyna
Copy link
Copy Markdown
Contributor Author

Please confirm if the provider_endpoints_support.json changes are intentional

Yes, intentional since sagemaker provider was added.

I've removed the provider_endpoints_support.json update and only included doc changes.

@joereyna
Copy link
Copy Markdown
Contributor Author

@greptile-apps review this pr

@ghost
Copy link
Copy Markdown

ghost commented Mar 18, 2026

@joereyna can you confirm this diff is against v1.82.0-stable?

@ghost ghost merged commit b4a5e51 into BerriAI:main Mar 18, 2026
17 of 39 checks passed
This pull request was closed.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants