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

CohereCohere　·　当前代current generation Command A+Command A+（2026-05）

Command ACommand A模型家族谱系卡model family lineage card

Cohere 的企业向 agentic 家族：A+ 放开了 Apache-2.0（上一代还是门控），主打企业场景与多语言。尺寸在 218B 级，本地部署门槛高。本站未收其量化对比（GGUF 生态太薄，见下）。Cohere's enterprise agentic line: A+ opened up under Apache-2.0 (the previous generation was gated), pitched at enterprise and multilingual work. At 218B scale the local-deployment bar is high. This site carries no quantisation comparison (GGUF ecosystem too thin — see below).

本站未收 Command 的量化对比：command-a 系 GGUF 生态实测约 20k 下载（2026-10-30 HF 检索），远低于本站 ~30 万门槛——与 2026-10-09 盘点「主流不等于有量化版可用」同一现象。No quantisation comparison on this site: the command-a GGUF ecosystem measured ~20k downloads (HF search 2026-10-30), far below the ~300k threshold — the same 2026-10-09 finding that "mainstream does not mean quantisations people actually use".

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 2026-10-20

  • 出品属地Origin　国外出品foreign-made　出品方为 Cohere（加拿大）。Made by Cohere (Canada).
  • 国内下载Direct access　国内可直连direct OK　ModelScope 有 CohereLabs 官方组织镜像（2026-10-20 实测 200）；Hugging Face 直连不可达。Official CohereLabs org mirror on ModelScope (measured 200); Hugging Face direct unreachable.　·　国内镜像入口domestic mirror
  • 备案状态CAC filing　未见于备案名单not in filing list：公示名单未检索到 Cohere 相关条目（2026-10-20 检索 13 批附件）。No Cohere entry found in the public registry (13 batch attachments searched on 2026-10-20).

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

  • Command A+Command A+2026-05当前代current

    当前代：218B 总参 / 25B 激活 MoE（128 专家激活 8 + 1 共享），128K 输入 / 64K 输出上下文，支持视觉输入；Apache-2.0。Current generation: a 218B-total / 25B-active MoE (8 of 128 experts plus one shared), 128K input / 64K output context, vision inputs supported; Apache-2.0.

    官方页面 ↗official page ↗
  • Command A 系Command A line2025-07

    上一代：command-a-07-2025（HF 门控，本站 2026-10-20 直抓返回 401 需登录）与 vision / reasoning / translate 变体在册。Previous step: command-a-07-2025 (gated on HF — our 2026-10-20 fetch returned 401 without login) plus vision / reasoning / translate variants listed.

    官方页面 ↗official page ↗
  • c4ai-command-r 系c4ai-command-r line2024

    更早一代（command-r / r-plus / r7b 系列在册，CC 系非 OSI 许可）。The earlier line (command-r / r-plus / r7b listed, CC-family non-OSI licences).

    官方页面 ↗official page ↗

尺寸档Size tiers

  • 旗舰 218B MoE（25B 激活，多卡/大显存）Flagship 218B MoE (25B active; multi-GPU / large VRAM)
  • 上一代 r7b（7B 级，单卡）Previous r7b (7B-class, single card)

许可Licences

  • apache-2.0Apache-2.0（Command A+ README：「License: Apache 2.0」；上一代 command-a-07-2025 为门控发布）Apache-2.0 (Command A+ README: "License: Apache 2.0"; the previous command-a-07-2025 was gated) 许可原文 ↗licence text ↗

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

  • 官方 README：「25 billion active parameters and 218B total parameters … optimized for agentic, multilingual, and reasoning-heavy tasks」「Context length: 128K input」（企业向定位与基准数字为厂商口径）。Official README: "25 billion active parameters and 218B total parameters … optimized for agentic, multilingual, and reasoning-heavy tasks"; "Context length: 128K input" (the enterprise pitch and benchmark numbers are vendor framing).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：2026-10-30 · 本站不实测、不排名，未知就写未知。Verification snapshot: 2026-10-30 · no benchmarks, no rankings — unknown stays unknown.