跳到主要内容Skip to main content
投机取巧 Speculative Speculation

腾讯Tencent　·　当前代current generation Hy4 previewHy4 preview（2026-08）

腾讯混元Tencent Hunyuan模型家族谱系卡model family lineage card

中文开放权重家族里少有的「大总参 + 高激活」路线；Hy4 把上下文拉到 1M+，对长程 agent 与工程代码有明确指向。本站暂无 GGUF 量化收录，选型前先看官方许可与硬件门槛。An uncommon large-total/high-active open-weight Chinese family; Hy4 pushes context past 1M and targets long-horizon agents and engineering code. This site has no GGUF coverage yet — read the official licence and hardware floor before choosing.

本站暂无腾讯混元的 GGUF 量化收录（未达量化热度与可核验覆盖门槛）。No GGUF coverage for Tencent Hunyuan on this site yet (below the quantisation-heat and verifiable-coverage threshold).

先说清楚：这个站的数据怎么来的First, where this site's data comes from

✅ 实测字段Measured fields

直接从 Hugging Face 公开 API 读取的客观数据：文件体积（字节）、上下文长度、架构、许可、下载量、更新时间。Objective data read straight from the Hugging Face public API: file sizes in bytes, context length, architecture, licence, downloads, last updated.

✅ 推导Derived

由实测字段按公开规则算出：每权重比特数（依 llama.cpp 命名标准）、档位归属、显存需求区间。推导规则在Computed from the measured fields by published rules: bits per weight (per the llama.cpp naming convention), tier assignment, VRAM requirement range. The rules are fully published in METHODOLOGY.md 完整公开，任何人可复算。 — anyone can recompute them.

❌ 我们不说的What we do not claim

哪个量化质量更高、哪个更快、哪个模型更强——这些需要实测，本站不跑 benchmark，所以不装作知道。凡我们给出的主观判断，一律显式标注「我们的口径」。Which quantisation is higher quality, which is faster, which model is stronger — these need measurement. This site does not run benchmarks, so it does not pretend to know. Every subjective judgement we do offer is explicitly marked as our own reading.

本站This site does 不下载、不托管任何模型权重文件not download or host any model weight files。所有数字都来自公开元数据，点击即可回到发布页核对。. Every figure comes from public metadata, and clicking it takes you back to the release page to verify. 采集时间Collected: 2026-10-30（数据快照，非实时） (a snapshot, not live)

中文生态China availability 国产 / 国内可直接用 / 备案状态 · 快照domestic / direct access / CAC filing · snapshot 2026-11-03

  • 出品属地Origin　国产domestic　出品方为腾讯（深圳）。Made by Tencent (Shenzhen).
  • 国内下载Direct access　国内可直连direct OK　hunyuan.tencent.com 官方页大陆直连可达（2026-11-03 未逐条复测）。Official hunyuan.tencent.com page reachable from the mainland (not re-tested item by item on 2026-11-03).　·　国内镜像入口domestic mirror
  • 备案状态CAC filing　备案未核验filing unknown：腾讯混元相关备案条目未逐条核验；以网信办公示名单为准。Filing entries for Tencent Hunyuan not verified item by item; the CAC public registry is authoritative.

备案信息以网信办《生成式人工智能服务已备案信息》公示名单为准（按批次滚动更新，本站只抄编号不解读）；「国内可直连」是本站从大陆网络直连实测的结果，不是厂商承诺。Filing status is quoted from the CAC public registry (updated batch by batch; this site copies the number and does not interpret it). "Direct access" is this site's own reachability measurement from a mainland network, not a vendor promise.

代次Generations

  • Hy4 previewHy4 preview2026-08当前代current

    当前代旗舰：770B 总参 MoE、49B 激活、上下文超过 100 万 token，官方主打长程 agent、编码与办公任务。Current flagship: a 770B-total MoE with 49B active parameters and a context beyond one million tokens, positioned for long-horizon agents, coding and office work.

    官方页面 ↗official page ↗
  • Hy3 / Hy3 previewHy3 / Hy3 preview2026-07

    Hy3 于 2026-07 正式发布；preview 于 2026-04-23 发布并开源：295B 总参、21B 激活、最大 256K 上下文，快慢思考融合 MoE。Hy3 launched officially in July 2026; the preview was released and open-sourced on 2026-04-23: 295B total, 21B active, up to 256K context, a hybrid fast/slow-thinking MoE.

    官方页面 ↗official page ↗

尺寸档Size tiers

  • Hy4 preview：770B 总参（MoE，49B 激活），1M+ 上下文Hy4 preview: 770B total (MoE, 49B active), 1M+ context
  • Hy3：295B 总参（MoE，21B 激活），256K 上下文Hy3: 295B total (MoE, 21B active), 256K context

许可Licences

  • other非宽松许可：使用前须读腾讯混元发布页的许可原文，本站不解读条款。Non-permissive: read the licence text on the Tencent Hunyuan release page; this site does not interpret it. 许可原文 ↗licence text ↗

官方能力声明（带出处）Official capability claims (with sources)

  • 腾讯官方：Hy4 preview 为 770B 总参、49B 激活、上下文超过 100 万 token，主打长程 agent、编码与办公。Tencent official: Hy4 preview has 770B total parameters, 49B active and a context beyond 1M tokens, pitched at long-horizon agents, coding and office work.出处 ↗source ↗
  • 腾讯官方：Hy3 preview 为 295B 总参、21B 激活、最大 256K 上下文，快慢思考融合 MoE。Tencent official: Hy3 preview has 295B total parameters, 21B active and up to 256K context, a hybrid fast/slow-thinking MoE.出处 ↗source ↗

以上均为厂商官方页面原文或官方自评基准的转述；涉及横评结论的，本站不背书——本站不跑 benchmark。The above are quotations of official pages or transcriptions of vendor-run benchmarks; this site does not endorse cross-model conclusions — it runs no benchmarks.

出处Sources

核验快照：2026-11-03 · 本站不实测、不排名，未知就写未知。Verification snapshot: 2026-11-03 · no benchmarks, no rankings — unknown stays unknown.