Meta (Llama - hosted)

v0.1.4rc1 Release Candidate

Release version 0.1.4rc1

Date not specified

v0.1.4.dev3 Development Release

Release version 0.1.4.dev3

Date not specified

v0.1.4rc2 Release Candidate

Release version 0.1.4rc2

Date not specified

v0.1.4rc3 Release Candidate

Release version 0.1.4rc3

Date not specified

Tool Call Encoding Bug Fixes - v0.1.4 Release

What's Changed fix: do not use python_tag when encoding non-code_interpreter tool_calls by @ehhuang in #283 fix: tool_call was not encoded by @ehhuang in #284 Full Changelog : v0.1.3...v0.1.4

Date not specified

v0.1.5rc1 Release Candidate

Release version 0.1.5rc1

Date not specified

Software v0.1.5rc2 Release

Release version 0.1.5rc2

Date not specified

New Sage — Signal Interpreter Release: v0.1.5rc3

Release version 0.1.5rc3

Date not specified

Product v0.1.5 Release

Release version 0.1.5

Date not specified

Llama.ai releases v0.2.0 with Llama 4 support

Llama 4 Support ( https://www.llama.com )

Date not specified

Meta Llama releases new rerank models and API changes

Meta Llama has added new rerank models with corresponding API changes, along with a demo script for easier implementation. Additional improvements include container optimization and conversation fixes. These are feature additions that enhance functionality without requiring immediate action from users.

Date not specified

Llama 4 Support Removed from Documentation — Verify Continued Support

Documentation for Llama 4 support has been removed from the README, indicating potential deprecation or discontinuation of Llama 4 features. Users currently relying on Llama 4 functionality should verify continued support and consider migration plans. Additional fixes were included for inference configuration handling.

Date not specified

Meta Llama v0.3.4 - Bug fixes and dependency updates

Minor release containing bug fixes for uninitialized write queue and adapter policies, plus Starlette version bump. These are maintenance updates that improve stability without requiring user action.

Date not specified

Model Filtering, Vector Store, and SQLite Bug Fixes Released

This release includes several bug fixes addressing model configuration filtering, vector store data persistence across server restarts, and SQLite functionality improvements. These are maintenance updates that improve system reliability and functionality without requiring user action.

Date not specified

UV Index Strategy & CI Shell Fixes - Technical Maintenance Update

Minor bug fixes addressing UV_INDEX_STRATEGY configuration issues and CI shell export problems. These are technical maintenance updates that improve system reliability without requiring user action.

Date not specified

Meta Llama v0.3.1 patch release — stability improvements

Minor patch release addressing continuous integration issues and including selective fixes from previous versions. This is a maintenance update that improves stability without introducing breaking changes or requiring user action.

Date not specified

Llama Stack introduces stable OpenAI-compatible APIs

Meta has restructured Llama Stack APIs with clear versioning (stable /v1/, experimental /v1alpha/ and /v1beta/, deprecated APIs). The update includes enhanced OpenAI compatibility with extra functionality support and comprehensive documentation improvements. This is primarily an organizational enhancement that improves API clarity and usability for developers.

Date not specified

Meta Llama: Documentation overhaul and API improvements

Meta Llama has migrated documentation to Docusaurus with modern formatting and standardized Ollama and Fireworks providers with OpenAI compatibility layer. The update includes improved model discovery, code refactoring, and promotes post_training and eval APIs to v1alpha status. These are informational improvements that enhance usability without requiring immediate action.

Date not specified

Meta Llama introduces unified test configuration and enhanced Together provider support

Meta Llama has implemented a new unified setups system for test configuration and enhanced the Together provider with embedding and dynamic model support. These updates also include automatic inference store setup during builds and a proposed API leveling framework. These are development improvements that enhance functionality without requiring immediate user action.

Date not specified

Meta Llama: Infrastructure Improvements, OpenAI Prompts API, RAG Tool Updates

Meta Llama added new testing infrastructure improvements including backwards compatibility tests for core APIs and integration of OpenAI Prompts API. The RAG Tool was updated to use Files API and Vector Stores API, along with improved error messaging for MCP server connections. These are enhancement updates that don't require immediate user action.

Date not specified