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

Mistral AI（法国）Mistral AI (France)　·　当前代current generation Small 4 / Medium 3.5Small 4 / Medium 3.5（2026-03）

MistralMistral模型家族谱系卡model family lineage card

欧洲唯一全尺寸开放权重厂商。特点：开源杯（Small）与旗舰杯（Medium）许可不同——Small 4 是 Apache-2.0，Medium 3.5 是改版 MIT；下载前按档位读许可。Europe's only full-spectrum open-weight vendor. The open tier (Small) and flagship tier (Medium) carry different licences — Small 4 is Apache-2.0, Medium 3.5 is Modified MIT. Read the licence per tier before downloading.

本站未收录 Mistral 系列的量化对比：Mistral-Small-4-119B 的第三方 GGUF 下载量约 8.4 万（2026-10-30 复测），未过 METHODOLOGY 的量化生态门槛（约 30 万）。This site does not cover Mistral quantisations: Mistral-Small-4-119B's third-party GGUF downloads (~84k, re-measured 2026-10-30) fall below the METHODOLOGY ecosystem threshold (~300k).

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 Mistral AI（法国）。Made by Mistral AI (France).
  • 国内下载Direct access　直连未核验direct unknown　ModelScope 站内检索 Mistral 未命中官方镜像（2026-10-20）；Hugging Face 直连不可达——同 llama 口径记 unknown。No official Mistral mirror found on ModelScope (searched 2026-10-20); Hugging Face direct unreachable — unknown for the same reason as Llama.
  • 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 Mistral 相关条目（2026-10-20）。No Mistral entry found in the public registry (searched 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.

  • 网信办《生成式人工智能服务已备案信息》滚动公告页（13 个批次附件 PDF，2026-10-20 全量下载检索）：CAC · filed generative-AI service information (rolling announcement page, 13 batch attachment PDFs, full download & search on 2026-10-20): https://www.cac.gov.cn/2024-04/02/c_1713729983803145.htm

代次Generations

  • Mistral-Small-4-119B-2603Mistral-Small-4-119B-26032026-01

    当前代开源杯：119B MoE（每 token 激活 6.5B）、256k 上下文、多模态输入、Apache-2.0。Current open tier: 119B MoE (6.5B activated per token), 256k context, multimodal input, Apache-2.0.

    官方页面 ↗official page ↗
  • Mistral-Medium-3.5-128BMistral-Medium-3.5-128B2026-03

    当前代旗舰（非开源杯）：官方定位「first flagship merged model」，密集 128B、256k 上下文，Modified MIT 许可。Current flagship tier: officially positioned as the "first flagship merged model", dense 128B with 256k context, under a Modified MIT licence.

    官方页面 ↗official page ↗
  • Mistral-Small-3.2-24BMistral-Small-3.2-24B2025-06

    上一代开源杯 24B，官方口径延迟比 3.1 降 40%。Previous open tier at 24B; officially 40% lower latency than 3.1.

    官方页面 ↗official page ↗
  • MagistralMagistral2025-06

    推理线（ Magistral Small/Medium ），Medium 3.5 官方文档说它在 Le Chat 里被 3.5 取代。The reasoning line (Magistral Small/Medium); per the official Medium 3.5 documentation, it is superseded by 3.5 in Le Chat.

    官方页面 ↗official page ↗

尺寸档Size tiers

  • Small 24B–119B（MoE）：本地可跑的开源杯Small 24B-119B (MoE): the locally runnable open tier
  • Medium 128B：旗舰密集杯Medium 128B: flagship dense tier

许可Licences

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

  • 官方 README（Small 4）：「119B parameters, with 6.5B activated per token」+「256k context length」+ 128 experts / 4 active 的 MoE 结构。Official README (Small 4): "119B parameters, with 6.5B activated per token", "256k context length", MoE with 128 experts and 4 active.出处 ↗source ↗
  • 官方 README（Medium 3.5）：「a dense 128B model with a 256k context window」，指令遵循、推理、编码合一；取代 Medium 3.1 与 Magistral。Official README (Medium 3.5): "a dense 128B model with a 256k context window", unifying instruction-following, reasoning and coding; supersedes Medium 3.1 and Magistral.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.