Tencent Hy4 flagship lands on Together AI with 1M context, native vision and computer use
togethercompute · x · 2026-11-20
Together AI now serves Tencent's flagship Hy4 model on serverless inference at $0.83/1M input, $2.50/1M output and $0.042/1M cached tokens.
Key points:
- Hy4 is the successor to Hy4 Preview and the first Hunyuan model with native vision: it reads charts and documents, reasons across multiple images, and solves visually grounded math problems
- Ships with a 1M-token context window, adjustable reasoning effort, tool calling, and structured output
- Built for long-horizon agentic work: operates both CLI tools and GUIs, driving professional software like Blender and CAD across multi-step, multi-app tasks
- Trained against real workflows in office work, financial analysis, and game development (Unreal/Unity/Godot), with reportedly large reductions in speculative errors vs. the preview
More from Models
- Together tests Tencent Hy4: exact chart-to-JSON extraction and a one-prompt browser game — togethercompute · 2026-11-20
- Anthropic Gets Claude Prefill Running on AMD MI355X in Three Days — AccBalanced · 2026-11-20
- Meta's AI Assistant Muse Read Private Signal Chats Without Permission, Reuters Finds — matthew_d_green · 2026-11-20
- Vibe-coded 'Photoshop killers' break on a simple PSD, only Photopea renders it — ssh4net · 2026-11-20
- Microsoft's MAI Code 2 lands in Copilot, early users say it matches Claude Opus — AmyKateNicho · 2026-11-20
- Meta's new model reportedly beats ChatGPT, Claude and Gemini on a key benchmark — Dense_Primary629 · 2026-11-20