vllm Releases
139 releases of vllm-project/vllm
- v0.5.4
A significant stability update that hardens sandbox lifecycle management, optimizes API server interactions, and introduces scoped-token authorization and sandbox recycling in the Python SDK.
Jul 30, 2026
- v0.26.0
A major release introducing the Inkling model family, DeepSeek-V4 performance optimizations, and matured KV offloading capabilities.
Jul 27, 2026
- v0.25.1
A patch release containing targeted bug fixes for TorchCodec and mixed-dtype quantization issues.
Jul 14, 2026
- v0.25.0
A major release making Model Runner V2 the default execution path, removing PagedAttention, and introducing the Streaming Parser Engine.
Jul 11, 2026
- v0.24.0
A major release featuring MiniMax-M3 support, DeepSeek-V4 optimizations, Model Runner V2 support for quantized models, and DiffusionGemma.
Jun 29, 2026
- v0.23.0
A major release featuring DeepSeek-V4 maturity, Model Runner V2 expansion, Rust frontend growth, and Gemma 4 support.
Jun 15, 2026
- v0.1.464-betaDiffusionGemma + Gemma 4 MTP
Adds support for DiffusionGemma, Gemma 4 MTP, and MiniMax-M3, introduces a new Model Hub and Chat with Files/RAG features, and enhances hardware support and tool calling reliability.
Jun 12, 2026
- v0.1.451-betaGemma 4 MTP + Bug Fixes
Focuses on bug fixes, cross-platform support improvements, and enabling audio input for Gemma 4 GGUFs, including UI refinements for tool calling and API compliance.
Jun 10, 2026
- v2026.6.5Hermes Agent v0.16.0 (2026.6.5) — The Surface Release
This release introduces a brand-new native desktop application and a comprehensive web dashboard admin panel, alongside a trimmed default skill set and Simplified Chinese support.
Jun 6, 2026
- v0.22.1
A patch release with bug fixes, new model support for JetBrains Mellum v2, and AMD Zen CPU acceleration.
Jun 5, 2026
- v0.1.44-betaGemma 4 12B, New UI, MCP, Projects
A beta release for Unsloth focusing on MCP integration, Projects, Canvas, and the new chat UI, alongside Gemma 4 12B support.
Jun 3, 2026
- v0.1.43-betaCUDA 13.3, Windows, Mac update
A beta update for Unsloth focusing on hardware support (Apple Silicon, Windows, CUDA 13.3) and MCP enhancements.
May 31, 2026
- v2026.5.29.2Hermes Agent v0.15.2 (2026.5.29.2)
A minor patch release for Hermes Agent focusing on packaging fixes for bundled plugin manifests.
May 29, 2026
- v0.22.0
A major release featuring DeepSeek V4 maturity, Model Runner V2 default for Qwen3, Rust frontend integration, and batch-invariant performance improvements.
May 29, 2026
- v2026.5.29Hermes Agent v0.15.1 (2026.5.29) — The Patch Release
A hotfix patch release addressing dashboard infinite reload loops, Docker configuration changes, and various Kanban/Worker improvements.
May 29, 2026
- v2026.5.28Hermes Agent v0.15.0 (2026.5.28) — The Velocity Release
The 'Velocity Release' introducing a massive code refactor, a new multi-agent Kanban platform, and significant performance improvements.
May 28, 2026
- v0.1.42-betaAn Update before Revamp!
An update before a major revamp introducing API connections, external backend support, and significant security improvements.
May 26, 2026
- v0.73.10.73.1
A bug fix release for fzf addressing environment variable handling and HTTP body accumulation issues in server mode.
May 25, 2026
- v0.73.00.73.0
A feature release for fzf adding Nushell integration, idle timers, and improved UI positioning options.
May 23, 2026
- v2026.5.16Hermes Agent v0.14.0 (2026.5.16)
A major foundation release that introduces PyPI installation, new OAuth providers (Grok), Microsoft Teams integration, X search tool, and significant performance optimizations including debloating and faster cold starts.
May 16, 2026