Xinference

AI inference platform for private and scalable model deployment.

GitHub:

[Xinference @Xorbitsio]

🆕Meet Xagent, our new open-source agent system.

As models get smarter, workflows get in the way. Every requirement change breaks your graph.

🙌Xagent's approach: goals not flows. Describe what you want—planning and execution emerge naturally.👍

https://t.co/29KrDjhPCp

[Xinference @Xorbitsio]

What you get:
✅ Custom Discord role
✅ Contributor directory
✅ Beta access + event priority
✅ Swag drops
✅ Free Xinference paid product + Xagent Starter

5 contributions in one track → Senior Contributor (6 months free, private repo, direct line to the team).

Build something with Xinference. We’d love to see it🤩. → https://t.co/SHXYLhcH6U

[Xinference @Xorbitsio]

🔶 Introducing the Xinference Builders Program 🔶

🌱 Build with Xinference. Grow with the community.
One real contribution gets you in — a merged PR, a great bug report, an accepted feature idea, a tutorial, or a project you shipped on top of us.

👉 https://t.co/DOfSaUM8yv

[Xinference @Xorbitsio]

Five ways to contribute — badges & perks stack:
🛠 Code Contributor — merged PRs
💡 Idea Contributor — accepted feature requests
🐞 Bug Hunter — reproducible, fixed bugs
🧩 Ecosystem Creator — tools, integrations, tutorials
📣 Ecosystem Advocate — original content, anywhere public

[Xinference @Xorbitsio]

@NousResearch Big congrats!

[Xinference @Xorbitsio]

Xagent v0.4.2 is out 🚀

What’s new:

• Workforce multi-agent collaboration
• Runtime task delegation with full UI + public API
• In-browser .pptx preview
• Microsoft Graph MCP tools for Teams, Outlook & OneDrive
• Celery background tasks for more stable long-running jobs
• Email-based password reset

More Info: https://t.co/6ZgCLJQz9M

[Xinference @Xorbitsio]

@Lovable @tan_stack bravo!

[Xinference @Xorbitsio]

We just hit 300+ models in our system, and MiniMax M3 is already here. Hard to keep up with how fast AI is moving.

[MiniMax (official) @MiniMax_AI]

Introducing MiniMax M3: The First Open-Weights Model to Combine Three Frontier Capabilities

- Coding & Agentic Frontier: 59.0% SWE-Bench Pro, 66.0% Terminal Bench 2.1, 34.8% SWE-fficiency, 28.8% KernelBench Hard, 74.2% MCP Atlas
- MiniMax Sparse Attention scales context to 1M
- Natively Multimodal from Step Zero

API: https://t.co/fHRdSV7BwZ
Token Plan: https://t.co/BDCycxepZw
🚀New! MiniMax Code: https://t.co/GvB4YiB6Ul

Weights & Tech Report in ~10 Days

[Xinference @Xorbitsio]

Wake up, a new model just dropped. Exciting to see multimodal agents moving fast.

[Qwen @Alibaba_Qwen]

👏👏 Introducing Qwen3.7-Plus — a multimodal agent model that unifies vision and language into one versatile agent foundation.

✅ Multimodal interactive hybrid agent: unified GUI & CLI operation across visual and text tasks
✅ Versatile coding agent & productivity assistant with full-modality input
✅ Visual Agent: perception, reasoning, grounding, and search-augmented QA
✅ Cross-harness generalization across diverse agent frameworks

One model. Sees, thinks, codes, acts.🙌🙌

Now available via API on Alibaba Cloud Model Studio. Try it — let us know what you build.😎

🔗🔗⬇️⬇️
Blog：https://t.co/pVYf0h3NNa
Qwen Studio：https://t.co/HUYgFW4cYf
API：https://t.co/viL0cXrMzW

[Xinference @Xorbitsio]

@deepseek_ai That’s game over.

[Xinference @Xorbitsio]

@FireworksAI_HQ @FactoryAI Great to see open models getting more love and real production usage.

[Xinference @Xorbitsio]

@dify_ai Yes, this is exactly what we hear from teams too. The demo is often the easy part: production is where serving, monitoring, cost, and governance all start to matter.

[Xinference @Xorbitsio]

Xinference v2.9.0 is now available 🔥

This release brings new model support, stronger logs and monitoring, and early access to the next Web UI experience.

Highlights:
🐳 DeepSeek-V4 Flash & Pro support
⚙️ vLLM 0.21.0 support for DeepSeek-V4
📊 New Log Center UI with Elasticsearch integration
📈 Monitoring UI improvements
🔐 Experimental advanced auth
🎨 Experimental new Web UI
Install:
pip install 'xinference==2.9.0'

More Info: https://t.co/TRHdACNDhg

[Xinference @Xorbitsio]

「Xagent v0.4.1 is here 🚀」

This release focuses on faster responses, more stable task execution, and a smoother file + connector experience.

Highlights:
⚡ Final Answer Streaming
🧠 Redis hot-path cache
📁 Durable file storage with smarter file_id management
🧩 MCP visibility controls
📄 Auto document preview
🔑 SDK Backend Phase 1

Upgrade:
cd xagent
git pull
docker compose up -d

More Info: https://t.co/6ZgCLJR6Zk

[Xinference @Xorbitsio]

「Xagent v0.3.4 is out」

A stability-focused update for smoother MCP, KB / RAG, and Web Crawler workflows.

📚 More stable MCP tool loading.
⚡ Streamable HTTP Transport support.
🏃 Better KB / RAG creation, retrieval, upload, and LanceDB concurrency.
🌐 Browser TLS Impersonation for Web Crawler.
🤖 Fixes for React retry loops, Tool Call Schema parsing, and Custom API Preview behaviour.

👉Quick Start: https://t.co/0lo48JkECd
👉GitHub: https://t.co/6ZgCLJR6Zk

[Xinference @Xorbitsio]

🧑‍🚀Xinference v2.8.0 is out.🧑‍🚀

Running models in production just got easier to observe and more resilient to failures.

🤳What’s new:
• Monitor Cluster / Worker / Model with Prometheus Metrics
• Use the built-in Grafana Dashboard in Web UI
• Recover from worker model failures automatically
• Get stronger vLLM support with v0.19.0 and TTFT histogram metrics

🦿Upgrade:
pip install 'xinference==2.8.0'

Full release notes: https://t.co/sGNKPOM0jJ

[Xinference @Xorbitsio]

🚀 Xagent v0.3.2 is here! 🦾
Experience seamless MCP app integration, a secure Docker sandbox for code execution, and smarter 🙌web retrieval via Exa AI.
Plus, enjoy full custom API support for maximum flexibility. 🥳
Upgrade now: https://t.co/29KrDjhPCp

[Xinference @Xorbitsio]

🚀 Xinference v2.5.0 is out!
This update brings support for Gemma-4 & Qwen3-TTS , while the Web UI now lets you save and swap multiple deployment configs via localStorage. Reliability is also boosted with auto-recovery for model replicas after supervisor restarts. 🤖✨ #AI

[Xinference @Xorbitsio]

Xagent v0.2.2 is out! 🎙️
Now featuring Audio Model support and DuckDB integration for flexible data processing.
Enjoy enhanced sandbox configs, global image uploads, and seamless task searching.
Upgrade now! ✨https://t.co/29KrDjhPCp