Tencent launches flagship Hy4 model, open weights and system card promised soon
testingcatalog · x · 2026-11-04
Tencent released Hy4, its new flagship model, claiming top-tier open-model performance across benchmarks. It's live on Tencent Cloud, OpenRouter, CodeBuddy, WorkBuddy, and Hy AI Studio.
Key points:
- Open source plan: weights will be released openly, with ecosystem support partnered with the vLLM (Inferact Inc.) and SGLang communities
- Transparency: a comprehensive system card covering safety evaluations and technical details will ship with the weights
- Benchmarks (chart): across 12 evals including ProgramBench, FrontierCode 2.0, Terminal Bench 4.0, and OSWorld 2.1, Hy4 competes with GLM 5.3, DeepSeek-V4.1-Flash, Kimi K3.1 Pro, Claude Opus 5.5, and GPT 6.1 Sol — first-tier in most, though still behind Claude Opus 5.5 on coding/terminal tasks (e.g. 31.0 vs 46.4 on FrontierCode 2.0)
More from Models
- Hobbyist rewrites Qwen 0.6B weights with 39.8M tokens, no gradients, in 3h12m — Altruistic_Heat_9531 · 2026-11-04
- Same leak, reshared: ChatGPT's five family plan prices spotted again — heyshrutimishra · 2026-11-04
- KOL visits StepFun's Shanghai office, hints a powerful new model is coming — teortaxesTex · 2026-11-04
- Fable 5.5 delayed as Anthropic reportedly bets OpenAI's Astra 6.2 won't beat it — haider1 · 2026-11-04
- ByteDance open-sources Axion 1.0-35B agent foundation model with full reproducible recipe — pmttyji · 2026-11-04
- Community asks Moonshot when Kimi K3.1 model weights will drop — bgmshana · 2026-11-04