PulseAugur
中
实时 21:11:26
English(EN) Mistral Large 4 (Mistral AI) is open-weight with a 524.3k token context — longer than most commercial models, at $0.68/$2.09 per 1M. That makes it a strong cand

Mistral Large 4 发布,在网络安全领域挑战中国AI模型

Mistral AI 发布了 Mistral Large 4 (ML4),这是一款万亿参数的混合专家模型,旨在与美国和中国的领先AI模型竞争。Artificial Analysis 的独立测试表明,ML4 在网络安全方面表现强劲,在特定测试中优于一些中国模型,但在更广泛的智能基准测试中落后于一些中国开放权重模型。该模型在 Nvidia Grace Blackwell GPUs 上训练,提供 100 万 token 的上下文窗口和原生图像输入,预计将于 10 月底发布其权重。 AI

影响 使 Mistral AI 成为美国和中国AI实验室的有力竞争者,尤其是在网络安全等专业领域。

排序理由 前沿实验室模型发布,附带系统卡。

在 Mastodon — fosstodon.org 阅读 →

AI 生成摘要 · Google Gemini · 来自 12 个来源。 我们如何撰写摘要 →

Mistral Large 4 发布,在网络安全领域挑战中国AI模型

本文如何被排名

Signal score
0 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Frontier Release
前沿实验室模型发布,附带系统卡。
Source corroboration
12 independent sources
Strong cross-source corroboration — multiple independent publishers covered this within the clustering window.
Topics
model release, product
Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance
High
Clearly on-topic for AI-industry coverage.
Story freshness
46 days old
Aged out of breaking-news scoring windows; ranking reflects the durable signal from the full source set.

完整方法见我们的编辑标准。

报道来源 [12]

  1. Tom's Hardware TIER_1 English(EN) · Shane Downing ·

    Mistral 的新 Large 4 在独立测试中落后于一些中国开源模型

    Artificial Analysis scores Mistral’s new Large 4 at 38, behind open models from Xiaomi, Z.ai, Moonshot, and DeepSeek.

  2. MarkTechPost TIER_1 English(EN) · Asif Razzaq ·

    Mistral AI 发布 Mistral Large 4 (Le Chonk):一款 1.05T 参数的多模态 MoE 模型

    <p>Mistral AI has released Mistral Large 4, nicknamed Le Chonk, as a public preview. It is a 1.05 trillion parameter Mixture of Experts model with 49 billion active parameters, native image input, and a 1 million token context window, trained on 3,800 NVIDIA Grace Blackwell GPUs …

  3. Mastodon — fosstodon.org TIER_1 Italiano(IT) · [email protected] ·

    🧠 Mistral Large 4 融合视觉、推理和工具。训练后,模型将在可验证的尝试和长工作流上训练代理。API 预览已上线

    🧠 # Mistral Large 4 combina visione, ragionamento e strumenti. Il post-training allena gli agenti su tentativi verificabili e workflow lunghi. Preview API disponibile, pesi annunciati entro fine ottobre. 👉 I dettagli: https:// lnkd.in/p/drQ6rgvu ] ___ ✉️ 𝗦𝗲 𝘃𝘂𝗼𝗶 𝗿𝗶𝗺𝗮𝗻𝗲𝗿𝗲 𝗮𝗴𝗴𝗶𝗼𝗿𝗻𝗮…

  4. Mastodon — fosstodon.org TIER_1 English(EN) · [email protected] ·

    Mistral Large 4 (Mistral AI) 开放权重,拥有 524.3k token 上下文 — 比大多数商业模型更长,价格为 0.68 美元/2.09 美元每百万 token。这使其成为一个强有力的竞争者

    Mistral Large 4 (Mistral AI) is open-weight with a 524.3k token context — longer than most commercial models, at $0.68/$2.09 per 1M. That makes it a strong candidate for long-document reasoning tasks without vendor lock-in. https:// olud.ai/latest.html # AI # LLM # OpenSource

  5. Mastodon — mastodon.social TIER_1 English(EN) · Synopsis360 ·

    Mistral AI 发布 Large 4 — 欧洲在 AI 竞赛中的回应。它在网络安全方面优于一些中国模型,并在编码、金融和...

    Mistral AI just unveiled Large 4 — Europe's answer to the AI race. It beats some Chinese models in cybersecurity and is closing the gap in coding, finance & more. Drops publicly Oct 27. Wild detail: it tried to go beyond its test environment during development. Mistral says it wa…

  6. Mastodon — mastodon.social TIER_1 Türkçe(TR) · berktech ·

    🤖 AI新闻 Mistral Large 4 推出:欧洲AI模型Mistral AI的新模型Mistral Large 4拥有1万亿参数,1万亿

    🤖 Yapay Zeka Haberi Mistral Large 4 Tanıtıldı: 1 Trilyon Parametreli Avrupalı Yapay Zeka Modeli Mistral AI'ın yeni modeli Mistral Large 4 özellikleri, 1 trilyon parametre yapısı ve yasal süreçlerdeki başarısıyla öne çıkıyor. Detaylar haberimizde. 🔗 https:// birolkurnaz.com/l/0aCi…

  7. Mastodon — mastodon.social TIER_1 English(EN) · strike007 ·

    Mistral Large 4 推出,采用 1.05T 参数 MoE 模型,拥有 49B 活跃参数、原生图像输入和 1M token 上下文窗口,使用 3,800 NVI 训练

    Mistral Large 4 launches as a 1.05T parameter MoE model featuring 49B active parameters, native image input, and a 1M token context window, trained on 3,800 NVIDIA Grace Blackwell GPUs. Open weights release later this month. # AI # MachineLearning # Mistral

  8. Mastodon — mastodon.social TIER_1 English(EN) · [email protected] ·

    📰 Mistral 发布号称可媲美中国最佳开源系统的新AI模型“Le Chonk” Mistral 发布了其1万亿参数的Mistral Large 4,或称“le Chonk”

    📰 Mistral Unveils New 'Le Chonk' AI Model It Says Rivals Best Open Systems From China Mistral has unveiled Mistral Large 4, or "le Chonk," a 1-trillion-parameter model that it says is the strongest open-weight AI model developed outside China by a wide margin. The model is enteri…

  9. Mastodon — mastodon.social TIER_1 Türkçe(TR) · berktech ·

    🤖 AI新闻 Mistral AI 宣布 Mistral Large 4,目标超越竞争对手模型。法国 AI 实验室 Mistral AI,同时也是美国 h

    🤖 Yapay Zeka Haberi Mistral AI, rakip modelleri geride bırakmayı hedefleyen Mistral Large 4'ü duyurdu Fransız yapay zeka laboratuvarı Mistral AI, hem Amerikan hem de Çin rakiplerini performansla aşmayı amaçlayan yeni büyük çok modelli (metin ve görsel işleyen) modeli Mistral Larg…

  10. Mastodon — mastodon.social TIER_1 English(EN) · [email protected] ·

    Mistral发布Mistral Large 4,一款旨在超越美国和中国竞争对手的新型大型多模态模型。这家法国AI实验室正将其

    Mistral has released Mistral Large 4, a new large multimodal model that aims to leapfrog both American and Chinese rivals. The French AI lab is positioning its latest model as a competitive alternative to leading foundation models from OpenAI, Anthropic and others. https:// techc…

  11. Mastodon — mastodon.social TIER_1 Türkçe(TR) · berktech ·

    🤖 AI新闻 Mistral AI推出拥有1万亿参数的Mistral Large 4,这是其迄今为止最大、最强大的模型

    🤖 Yapay Zeka Haberi Mistral AI, 1 trilyon parametreli Mistral Large 4'ü tanıttı Fransa merkezli yapay zeka girişimi Mistral AI, bugüne kadarki en büyük ve yetenekli modeli olarak konumlandırdığı Mistral Large 4'ü (ML4) tanıttı. Şirket içinde “Le Chonk” olarak adlandırılan model, …

  12. Mastodon — mastodon.social TIER_1 English(EN) · [email protected] ·

    Mistral AI 推出了其大型多模态模型 Large 4,代号为 le Chonk,拥有万亿参数(活跃 490 亿)。该公司使用 3,800 个 NVIDIA

    Mistral AI has launched its Large 4 multimodal model, nicknamed le Chonk, with one trillion parameters (49 billion active.) The company trained it on 3,800 NVIDIA GPUs in its EU datacentres and reports an 82% score on a vulnerability reproduction-and-patching test. How closely do…