Mistral Large 4（Le Chonk）现已开放权重 | Mistral

我们现已公开发布 Le Chonk 的权重，推动开放模型在智能体编程和网络安全领域的发展前沿。We release the weights of Le Chonk, advancing the frontier of open models in agentic coding and cybersecurity.

本月早些时候，我们公开预览了 Mistral Large 4（Le Chonk），这是一款拥有 1 万亿参数、原生支持多模态的旗舰模型。预览版的反响超出了我们的预期，我们感谢开发者社区、早期体验客户和合作伙伴给予的支持、协作与反馈。Earlier this month, we released Mistral Large 4 (Le Chonk), our flagship 1-trillion-parameter, natively multimodal model, in public preview. The response to the preview has exceeded our expectations, and we're grateful to our developer community, early-access customers, and partners for their support, collaboration, and feedback.

此后，我们在持续扩大强化学习规模的同时，也将这些反馈付诸实践，在编程和防御性网络安全方面取得了巨大进步。进步仍在加速，丝毫没有显现出饱和迹象。Since then, we've put that feedback to work while continuing to scale reinforcement learning, driving massive improvements in coding and defensive cybersecurity. Gains continue to accelerate, with no signs of saturation.

今天，我们发布这款改进版 Mistral Large 4 的权重。新模型在推理效率方面也有显著提升，让开发者能够获得更高性能，同时降低推理成本。Today, we're releasing the weights of this improved Mistral Large 4. The new model also features substantial gains in reasoning efficiency that give developers greater performance while lowering inference costs.

Mistral Large 4 是编程领域最出色的开放权重模型，也是全球整体能力最强的开放模型之一。它基于欧洲基础设施从头训练，拥有从诞生到发布全程清晰可溯的谱系，为组织构建可信赖的主权 AI 奠定了基础。Mistral Large 4 ranks as the best open-weight model for coding and among the world's most capable open models overall. Trained from scratch on European infrastructure with clear lineage from inception to release, it gives organizations a trusted foundation for Sovereign AI

Mistral 正在定义主权 AI 领域，并引领开放前沿。我们致力于打造全球能力最强的模型，并以开放方式发布，让开发者、企业和机构能够自主拥有、定制和掌控自己的 AI。Mistral is defining the sovereign AI category and leading the open frontier. We're committed to building the world's most capable models and releasing them openly, giving developers, enterprises, and institutions the power to own, customize, and control their AI.

为建设者而造。Built for the Builders.

Mistral Large 4 开启公开预览时，便在多项领先的代码基准测试中取得了优异成绩。此后仅经过三周的额外强化学习训练，它便跃居开放权重代码模型的前沿，在 FrontierCode 和 Terminal-Bench 上拔得头筹，同时在 Artificial Analysis Coding Agent Index 上达到最佳开放权重模型的水平，且 token 效率显著更高。Mistral Large 4 entered public preview with strong results across leading coding benchmarks. Just three weeks of additional reinforcement learning have since propelled it to the forefront of open-weight coding, taking the lead on FrontierCode and Terminal-Bench while matching the best open-weight models on the Artificial Analysis Coding Agent Index with substantially greater token efficiency.

在 Cognition 的专有 FrontierCode 2.0 基准测试中，Mistral Large 4 以 19.7% 的得分领先。该基准测试评估端到端代码质量，以及代码仓库维护者是否会合并智能体提交的更改。On Cognition's private FrontierCode 2.0 benchmark, which assesses end-to-end code quality and whether a repository maintainer would merge the agent's change, Mistral Large 4 leads with a score of 19.7%.

拓展智能体编程的帕累托前沿。Pushing the pareto frontier for agentic coding.

Mistral Large 4 以显著更少的 token 实现了前沿水平的编程性能，让大规模运行复杂的编程智能体更加经济实惠。Mistral Large 4 delivers frontier coding performance with significantly fewer tokens, making sophisticated coding agents more economical to run at scale.

在 Artificial Analysis Coding Agent Index 上，Mistral Large 4 的得分为 48%，与领先的开放权重模型 GLM-5.3 基本持平，同时平均每项任务使用的 token 少 47%。On the Artificial Analysis Coding Agent Index, Mistral Large 4 scores 48%, effectively tied with the leading open-weight model, GLM-5.3, while using 47% fewer tokens per task on average.

这些提升也体现在基于终端的编码工作流中。在 Terminal-Bench 4.0 上，Mistral Large 4 相较公开预览版提升了 13 个百分点，达到 41%，在开放权重模型中位居领先，同时将输出 token 数减少了 30%。The gains extend to terminal-based coding workflows. On Terminal-Bench 4.0, Mistral Large 4 improved by 13 percentage points to 41% compared to its public preview, taking the lead among open-weight models while reducing output tokens by 30%.

为了让开发者更好地权衡性能与成本，Mistral Large 4 现已支持三种推理模式：To give developers greater control over this tradeoff between performance and cost, Mistral Large 4 now supports three reasoning modes:

除了基准测试之外，Mistral Large 4 的编码能力也得到了人类偏好的验证。在 Surge Code 的盲测两两评估中，Mistral Large 4 优于 GLM-5.3、Kimi K3 和 DeepSeek V4.1 Flash 等领先的开放权重模型，在参与评估的模型中仅次于 Claude Opus 5.5。Beyond benchmarks, Mistral Large 4's coding capabilities are validated by human preference. In Surge Code's blind pairwise evaluations, Mistral Large 4 was preferred over leading open-weight models including GLM-5.3, Kimi K3, and DeepSeek V4.1 Flash, ranking behind only Claude Opus 5.5 among the models evaluated.

专为防守者打造。Built for the Defenders. 

自预览版发布以来，我们一直与领先的网络安全厂商合作，在真实环境中评估 Mistral Large 4，以测试其在漏洞复现和修补方面的优势。Since its preview, we've been working with leading cybersecurity vendors to evaluate Mistral Large 4 in real-world environments to test Mistral Large 4’s strengths in vulnerability reproduction and patching.

Mistral Large 4 在 Artificial Analysis 网络安全指数中得分 51%。该指数衡量模型发现代码库漏洞、复现并验证漏洞，以及在不破坏现有功能的前提下修补漏洞的能力。重要的是，Mistral Large 4 能够完成这些正当的防御性安全任务，而不会不必要地拒绝。Mistral Large 4 scores 51% on the Artificial Analysis Cyber Index, which measures a model’s ability to discover vulnerabilities in a codebase, reproduce and validate them, and patch them without breaking existing functionality. Crucially, Mistral Large 4 handles these legitimate defensive security tasks without unnecessary refusals.

安全性与能力同步提升。我们的模型内置防护机制，能够区分防御性网络安全任务与恶意攻击性网络安全任务。它在网络安全基准测试中仍表现出色，同时对恶意网络请求的平均拒绝率达到 96.6%，是我们测评的开放模型中最高的。Safety has improved alongside capability. Our model’s built-in guardrails distinguishes between defensive and malicious, offensive cybersecurity tasks. It retains strong performance on Cyber benchmarks, while achieving an average refusal rate of 96.6% on malicious cyber requests, the highest we measure among open models.

以下是全球网络安全领域的领军人物对 Mistral Large 4 的评价：Here’s what global leaders in cybersecurity have to say about Mistral Large 4:

Booz Allen 的测试表明，欧洲的 Mistral Large 4 模型在我们的源代码安全分析内部基准测试中表现出色。该模型在漏洞检测召回率方面能够媲美其他前沿网络安全模型，同时具备极具吸引力的成本优势。此外，Mistral 支持模型定制和数据主权部署，使我们能够根据任务需求以及对数据驻留有严格要求的客户环境进行定制。我们期待与 Mistral 合作，共同加强全球关键基础设施。Testing by Booz Allen demonstrated the success of the European Mistral Large 4 model against our internal benchmark for source code security analysis. The model was able to match vulnerability detection recall rates against other frontier cyber models with a compelling cost curve. And, Mistral's support for model customization and data-sovereign deployment also lets us tailor it for our mission needs and client environments with strict data-residency requirements. We look forward to working with Mistral to strengthen critical infrastructure around the globe.

布兹·艾伦·汉密尔顿国家网络安全业务总裁布拉德·梅代里Brad Medairy, President National Cyber at Booz Allen Hamilton

随着 AI 在安全运营中日益占据核心地位，将透明度和开放性视为优势的组织将更具优势。开放权重的网络安全模型让需要在自身控制下托管和定制 AI 的企业与政府也能享有这种透明度，同时使研究人员能够研究这些模型的行为，包括它们可能遭到滥用的情形。随着对手越来越多地利用 AI 加速行动，我们希望与 Mistral 等模型开发者携手，帮助安全团队充分运用这种掌控力，通过从设计之初就融入韧性，而非被动应对威胁，来奠定基础。Organizations that treat transparency and openness as strengths are positioned well as AI becomes central to security operations. Open-weight cybersecurity models extend that transparency to enterprises and governments that need to host and customize AI under their own control, and they allow researchers to study how these models behave, including where they could be misused. As adversaries increasingly use AI to accelerate their operations, we aim to work alongside model developers such as Mistral to help security teams build on that control, establishing a foundation for resilience by design rather than reactive threat management.

Amy Chang，思科人工智能威胁情报与安全研究负责人Amy Chang, Head of Cisco's AI Threat Intelligence and Security Research

当威胁以机器速度蔓延时，人类防御者根本无法跟上。更糟的是，无法由你掌控的防御并非防御，而是暴露风险。安全团队在掌握主动权时，才能发挥最佳表现。Mistral 的 ML4 让防御者能够审计架构、部署在自己的基础设施上，并完全掌握自主权。韧性始于自主，而开放权重模型有助于组织掌控局面，主导自身防御。When threats move at machine speed, human defense cannot keep pace. Worse, a defense you don’t control isn’t defense—it’s exposure. Security teams perform best when they take the steering wheel. Mistral's ML4 lets defenders audit the architecture, deploy on their own infrastructure, and keep full sovereignty. Resilience starts with autonomy, and open-weight models help give organizations the control to lead their own defense."

斯蒂芬妮·科恩，Cloudflare 首席战略官Stephanie Cohen, Chief Strategy Officer at Cloudflare

仅凭强大的 AI 模型，并不足以在网络安全领域占据优势。这需要深厚的安全专业知识、真实世界的威胁情报和严格的测试。我们与 Mistral Large 4 合作，测试前沿 AI 在推理复杂漏洞、加快修复进程方面的能力，并确保防御者的行动速度快于攻击者。Powerful AI models alone won't tip the scales in cybersecurity. It takes deep security expertise, real-world threat intelligence, and rigorous testing. Our work with Mistral Large 4 tests how well frontier AI can reason through complex vulnerabilities, accelerate remediation, and ensure defenders move faster than adversaries.

Bartley Richardson 博士，CrowdStrike 首席人工智能与自主系统官Dr. Bartley Richardson, Chief AI and Autonomous Systems Officer, CrowdStrike

Elastic 让安全团队能够将他们信赖的 AI 应用于自己掌控的数据，包括 Mistral 的开放模型；这些模型目前已在 Elastic Security 中获得支持。我们已在真实的网络安全工作负载中评估过 ML4。在这些场景中，模型需要对复杂数据进行推理，帮助分析师做出明智的调查和响应决策。这体现了我们正在建立的合作伙伴关系，旨在提供强大的开放权重模型，并结合客户所要求的安全上下文、透明度和控制能力。Elastic gives security teams the freedom to bring the AI they trust to the data they control, including Mistral’s open models, which are supported on Elastic Security today. We’ve already evaluated ML4 against real cybersecurity workloads, where models need to reason across complex data to help analysts make informed investigation and response decisions. This demonstrates the partnerships we are building to deliver powerful, open-weight models combined with the security context, transparency and control our customers demand.

詹姆斯·斯皮特里，Elastic 产品管理高级总监James Spiteri, Senior Director, Product Management, Elastic

ENISA 正在试验闭源和开放权重模型，以更好地了解它们的网络安全能力。开放权重模型使各组织能够掌控其韧性所依赖的技术。ENISA 已获得 Mistral Large 4 的使用权限，并正在对其进行测试，以识别安全漏洞、审计代码并分析防御日志。ENISA is experimenting with closed and open weight models in order to better understand their cyber capabilities. Open-weight models allow organisations to keep control of the technology their resilience depends on. ENISA access to and are testing Mistral Large 4 to identify security vulnerabilities, to audit code and analyse defensive logs.

ENISA传播部门负责人劳拉·赫文克Laura Heuvinck, Head of Communications of ENISA

我们相信，开放权重 AI 模型能够为组织提供更大的灵活性、选择空间和掌控力，从而在推进网络安全方面发挥重要作用。Mistral Large 4 等创新成果，展现了 AI 在变革网络防御方面的潜力，令人期待。我们认为，AI 持续创新是加强网络安全、帮助组织应对不断演变的威胁形势的重要机遇。We believe open-weight AI models have an important role to play in advancing cybersecurity by giving organizations greater flexibility, choice, and control. Innovations like Mistral Large 4 represent promising steps forward in AI's potential to transform cyber defense. We see continued AI innovation as an important opportunity to strengthen cybersecurity and help organizations navigate an evolving threat landscape.

Fortinet首席运营官约翰·惠特尔John Whittle, Chief Operating Officer at Fortinet

开放模型让安全研究人员和开发者能够检查、调整并基于 AI 进行构建。Mistral 正在拓展整个生态系统的构建能力，而我们在 NVIDIA 加速基础设施方面的合作，则帮助开发者将这些模型投入实际应用。Open models let security researchers and developers inspect, adapt and build on AI. Mistral is expanding what the ecosystem can build, and our collaboration on NVIDIA accelerated infrastructure helps developers put these models to work.

NVIDIA 开放模型与软件副总裁 Kari BriskiKari Briski, VP of Open Models and Software, NVIDIA

随着威胁变得越来越迅速、日益工业化，并如今实现自动化，防御也必须提升到新的规模。规模、速度、信任：这是 AI 时代网络安全面临的三大挑战。我们正与 Mistral AI 合作打造欧洲网络安全 AI，由我们的专家训练并在他们的掌控下运行。通过结合 AI 的强大能力、团队的专业知识以及我们的网络威胁情报，我们正在加快响应速度，从而以可信赖的欧洲技术，更全面、更迅速地提供保护。As threats become ever faster, industrialized and now automated, defense must operate at a different scale. Volume, speed, trust: these are the three challenges facing cybersecurity in the age of AI. With Mistral AI, we are building European cyber AI, trained by our experts and operated under their control. By combining the power of AI, the expertise of our teams and our Cyber Threat Intelligence, we are accelerating our response so that we can protect more broadly, more quickly and with trusted European technology.

Orange Cyberdefense 首席执行官于格·富隆Hugues Foulon, CEO of Orange Cyberdefense

Rubrik 支持 Mistral 以及对“主权 AI”的推动，为欧洲各国政府和受监管行业提供可信赖的选择，确保其数据不会被传输到其他国家。客户希望有多种模型可供选择，包括开放权重模型和私有模型，以满足数据驻留、延迟、成本和能力方面的要求。为落实我们增强网络韧性、加快从组织外部的 AI 驱动攻击及内部智能体攻击中恢复的工作重点，我们将继续对 Mistral Large 4 进行严格的网络安全测试，并密切合作，以满足不断变化的区域需求。Rubrik supports Mistral and the push for ‘sovereign AI,’ to give European governments and regulated industries a credible option that does not route their data to other countries. Customers want model choices, including open-weight and private models, to support data residency, latency, cost and capability requirements. As part of our focus to deliver cyber resilience and speed recovery from both AI-powered attacks outside an organization and internal agents, we’ll continue rigorous cybersecurity testing with Mistral Large 4 and work closely together to serve evolving regional demands.

Rubrik 首席产品官安妮卡·古普塔。Anneka Gupta, Chief Product Officer of Rubrik.

随着企业将 AI 投入生产环境，他们越来越希望能够自主掌控所使用的模型，并清楚了解模型行为、数据访问权限和智能体权限。Mistral Large 4 为开放权重 AI 带来了强大的网络安全能力和技术推理能力，许多欧洲企业已将其选为首选的主权模型。Zscaler Security for AI 产品现已集成 Mistral，并通过我们在欧洲的主权 Zero Trust Exchange 提供服务，帮助企业持续测试 AI 部署、梳理每个模型和智能体能够访问的资源，并在流量链路中实时执行策略。双方携手助力企业采用主权 AI，获得安全创新所需的可视性、控制力和数据保护。As enterprises bring AI into production, they increasingly want models they can run on their own terms, with clear visibility into model behavior, data access, and agent permissions. Mistral Large 4 brings strong cybersecurity and technical reasoning to open weight AI, and many European enterprises are already embracing it as their sovereign model of choice. Zscaler Security for AI offerings, delivered on our sovereign Zero Trust Exchange in Europe, now integrate with Mistral, helping organizations continuously test their AI deployments, map what every model and agent can reach, and enforce policy inline. Together, we are empowering enterprises to adopt sovereign AI with the visibility, control, and data protection they need to innovate securely.

Dhawal Sharma，Zscaler 人工智能与数据安全执行副总裁Dhawal Sharma, EVP of AI and Data Security, Zscaler

自预览版发布以来，我们一直在真实流量中密切监测与网络安全相关的使用情况。虽然我们观察到，该模型会拒绝配合针对真实目标的黑盒攻击，但我们认为它可能会助长真实攻击。我们正据此调整防护措施，并将在未来几个月继续更新。In the time since the preview, we carefully monitored cybersecurity-related use in our live traffic. Though we observed the model refusing to comply with black-box attacks against live targets, we see a potential for uplift to conduct real attacks. We are modifying our guardrails accordingly, and will continue updating them in the months to come. 

我们还向 Orange Cyberdefense 等企业级设计合作伙伴提供预览访问权限，让他们体验专为高级网络安全工作流优化的 Mistral Large 4 特定版本，涵盖渗透测试、漏洞评估和红队演练。访问须经过严格审核并遵守合同保障措施，测试仅限于受控环境中自有或明确获准使用的系统。We are also providing enterprise design partners such as Orange Cyberdefense, with preview access to a specialized variant of Mistral Large 4 optimized for advanced cybersecurity workflows, including penetration testing, vulnerability assessment, and red-teaming. Access is subject to rigorous vetting and contractual safeguards, with testing restricted to owned or explicitly authorized systems in controlled environments.

与此同时，客户正借助 Mistral Forge，使用专有安全数据、策略、检测逻辑和威胁环境来定制 Mistral Large 4。对于定义明确的网络安全任务，还可以将其进一步优化为更小、针对特定任务的变体，以提高准确性，让安全应用更加有效。Meanwhile, customers are tailoring Mistral Large 4 with proprietary security data, policies, detection logic, and threat environments using Mistral Forge. For well-defined cybersecurity tasks, it can be further optimized into smaller, task-specific variants for greater accuracy, making security applications more effective.

进一步了解用于网络安全的主权 AI，以及如何与我们合作。Learn more about sovereign AI for cybersecurity and how to work with us.

为工作而打造。Built for Work.

Mistral Large 4 在多项广泛的企业应用指标上持续取得进步。Mistral Large 4 has continued improving on a range of broad enterprise-use metrics. 

在金融相关工作方面，我们提升了模型执行真实工作流的能力，尤其是创建和编辑适用于金融和会计场景的电子表格。这体现在 Finch 得分升至 73.3%，在开放模型中仅次于 Kimi K3。
For finance-related work, we improved the model’s ability to perform real-world workflows, especially ones involving the creation and editing of spreadsheets for finance and accounting use cases. This is reflected in our Finch scores climbing to 73.3%, second only to Kimi K3 for open models.

网页浏览是大多数知识工作中至关重要的一环。我们的强化学习流程使 BrowseComp 得分达到 84.5%（起点是得分为 75% 的 SFT 检查点）。这使 Mistral Large 4 的表现超越了 GLM-5.3 和 DeepSeek Flash 4.1 1017 等模型。Web browsing is a critical part of a majority of knowledge work. Our RL pipeline enabled us to reach 84.5% on BrowseComp (starting from an SFT checkpoint that was achieving 75%). This brings Mistral Large 4 at a higher performance than models like GLM-5.3 or DeepSeek Flash 4.1 1017.

组织部署的智能体需要具备安全性。无论部署给内部成员还是外部用户，都不应让智能体使你暴露于系统提示词窃取或恶意代码注入等威胁之下。强化学习提升了我们在 B3 Agent Security Benchmark 上的成绩：Mistral Large 4 能抵御 97.1% 的攻击，而 Preview 版为 93.3%。这使其跻身开源模型前沿的顶尖水平，意味着任何托管该模型的人都能直接运行安全的智能体，无需额外调优。Agents deployed by an organization need to be secure. Whether deployed to internal members or external users, agents shouldn’t open you up to threats like system prompt exfiltration or malicious code injection. RL increased our gains on the B3 Agent Security Benchmark, with Mistral Large 4 resisting 97.1% of attacks vs 93.3% in Preview. This places it at the top of the OSS frontier, meaning anyone hosting the model can run a safe agent without additional tuning.

查看实际效果See it in Action

网络安全漏洞分诊：智能体读取收到的漏洞报告，调查源代码和产品背景，在隔离沙箱中复现漏洞，然后编写并迭代修补程序，直到金丝雀漏洞利用不再奏效。Cyber Vulnerability Triage: an agent reads incoming vulnerability reports, investigates the source code and product context, reproduces the flaws in an isolated sandbox, then writes and iterates on a patch until a canary exploit no longer works.

智能体式 Web 开发：无需人工协助，Mistral Large 4 构建了 Orbital Trace——一张实时 3D 地图，展示绕地球运行的大约 19,000 颗卫星和碎片。它获取真实的追踪数据，编写应用，在浏览器中运行应用以发现自身的错误和性能瓶颈，然后修复独立审查者指出的问题，直到满足所有要求。Agentic webdev: with no human help, Mistral Large 4 builds Orbital Trace, a live 3D map of roughly 19,000 satellites and debris pieces orbiting Earth. It pulls real tracking data, writes the app, uses it in a browser to catch its own bugs and slowdowns, then fixes what an independent reviewer flags until all requirements pass.

量子化学求解器：Mistral Large 4 从零开始编写软件，用来计算分子中电子的位置，并将计算结果与标准化学工具进行比对。随后，它围绕该软件构建了一个 3D 页面：选择苯或水等分子，拖动滑块将其拉开，观察化学键断裂时电子如何重新分布。Quantum chemistry solver: Mistral Large 4 writes, from scratch, software that calculates where electrons sit in a molecule, and checks its answers against a standard chemistry tool. It then builds a 3D page around it: pick a molecule such as benzene or water, drag a slider to pull it apart, and watch the electrons regroup as the bond breaks.

为主权而打造。Built for Sovereignty.

随着组织围绕自身知识和工作流程构建 AI，主权意味着它们能够掌控所依赖的模型以及所创造的智能。As organizations build AI around their knowledge and workflows, sovereignty means that they retain control of the models they depend on and the intelligence they create. 

借助 Mistral Large 4，这种掌控力始于模型本身。各组织能够检查、定制模型，并将其部署在自己的基础设施上。它们可以使用并改进 Mistral Large 4，而无需依赖 Mistral 通过 API 提供服务。Mistral Large 4 在欧洲从零开始开发和训练，让各组织清楚了解自己所依托技术的来源。With Mistral Large 4, that control begins with the model itself. Organizations have the ability to inspect, customize, and deploy the model on their own infrastructure. They can use and improve Mistral Large 4 without depending on Mistral to serve it through an API. Mistral Large 4 was developed and trained from scratch in Europe, giving organizations clarity about the origins of the technology they’re building on. 

控制权也延伸至 AI 在组织内部的运行方式。组织可以借助 Mistral Forge，围绕专有数据定制 Mistral Large 4；还可以使用 Mistral Studio 构建、管理并改进由此生成的智能体和应用。Mistral 的应用人工智能专家可以与客户团队协作，共同构建和部署这些定制系统。在这一过程中，团队能够了解所构建系统背后的架构和决策，以及训练和运行系统所用的方法，从而独立运行并改进自己的 AI。Control also extends to how AI operates inside an organization. Organizations can customize Mistral Large 4 around their proprietary data with Mistral Forge, and they can build, govern, and improve the resulting agents and applications using Mistral Studio. Mistral’s Applied AI experts can work alongside customer teams to build and deploy these tailored systems. Through that work, teams understand the architecture and decisions behind what they’ve built, along with the methods used to train and operate it, so they can run and improve their AI independently.

在 ABN AMRO，我们相信，开放透明的 AI 生态系统对于推动金融领域以负责任的方式创新至关重要。Mistral Large 4 等进展表明，主权 AI 模型和开放权重 AI 模型能够为欧洲组织拓展可信解决方案的选择范围。我们对目前取得的进展感到鼓舞，并将继续探索先进 AI 如何助力强化关键能力，包括网络安全。At ABN AMRO, we believe that open and transparent AI ecosystems are essential for driving innovation responsibly within the financial sector. Developments such as Mistral Large 4 demonstrate how sovereign and open-weight AI models can expand the range of trusted solutions available to European organizations. We are encouraged by the progress being made and continue to explore how advanced AI can contribute to strengthening critical capabilities, including cybersecurity.

Coen Klaver，ABN AMRO Bank N.V. 网络防御总监。Coen Klaver, Director Cyber Defense at ABN AMRO Bank N.V.

我们看到的 AI 模型快速进步，为芯片制造领域的先进工程带来了令人振奋的机遇。我们对 Mistral Large 4 的初步体验令人鼓舞，尤其适用于更高模型性能至关重要的场景。Mistral 对持续创新的投入，助力我们加快 AI 在整个半导体产品生命周期中的应用。The rapid progress we are seeing in AI models creates exciting opportunities for advanced engineering in chip manufacturing. Our early experience with Mistral Large 4 is promising, particularly for use cases where higher model performance matters. Mistral’s commitment to continued innovation supports our efforts to accelerate the application of AI throughout our semiconductor product lifecycle.

Arnaud Hubaux，人工智能项目与战略负责人。Arnaud Hubaux, Head of AI Program and Strategy.

Mistral Large 4 是金融级 AI 迈出的重要一步：这是一款开放权重模型，我们可以根据自身需求对其进行微调，部署在自有基础设施上，并将其紧密集成到数字渠道中。Hey Belfius 是这一合作如何帮助我们重新思考客户与银行互动方式的典范。Mistral Large 4 is an important step forward for financial‑grade AI: an open‑weight model we can fine‑tune to our needs, deploy on our own infrastructure and closely integrate into our digital channels. Hey Belfius is a flagship example of how this partnership helps us rethink how customers interact with their bank.

Nicolas Goosse，Belfius 数据洞察、分析与人工智能负责人Nicolas Goosse, Head of Data Insights, Analytics & AI at Belfius

掌握我们的人工智能，首先要掌握其架构和数据。对于最敏感的应用场景，能够使用欧洲开放权重模型并将其部署在自有基础设施上，是增强战略自主性和组织韧性的关键抓手。法国农业信贷银行人工智能与 Mistral 的产业合作，从集团层面应对了这些挑战。“Le Chonk”模型以其推理能力和在网络安全领域的出色表现脱颖而出，是这项合作所开启的种种可能性的一个令人期待的例证。Mastering our AI begins with mastering its architecture and data. For the most sensitive use cases, having access to European open-weight models, which we can deploy on our own infrastructure, is a vital lever for strengthening our strategic autonomy and the resilience of our organizations. The industrial partnership between Crédit Agricole Artificial Intelligence and Mistral addresses these challenges at the Group level. The "Le Chonk" model, which stands out for its reasoning capabilities and performance in the field of cybersecurity, is a promising illustration of the possibilities opened up by this partnership.

Oliver Biton，法国农业信贷银行集团首席信息官Oliver Biton, Group CIO, Crédit Agricole

我们对这款开放权重模型强大的科学与视觉能力感到振奋。我们的工作高度专业化，因此模型的适配能力至关重要。Mistral Large 4 为我们提供了一个基础，让我们能够围绕科学家和工程师处理地球观测数据的实际方式构建工具，涵盖从影像到技术文档的各种数据。We're excited about an open-weight model with strong scientific and vision capabilities. Our work is highly specialized, so the ability to adapt a model matters. With Mistral Large 4, we see a foundation for building tools around the way our scientists and engineers actually work with Earth observation data — from imagery to technical documents.

ESA Φ-lab负责人朱塞佩·博尔吉Giuseppe Borghi, Head of ESA Φ-lab

AI 主权并非在于选择调用哪个 API，而在于能够按照自己的方式运行、定制和构建 AI。Mistral Large 4 将强大的编程、多模态 AI 和网络安全能力开放出来，为开发者、企业和国家提供出色的开放基础，让他们能够构建自己的 AI，并抵御网络威胁。AI sovereignty isn't about choosing which API to call. It's about having the freedom to run, customize, and build AI on your own terms. Mistral Large 4 brings powerful capabilities in coding, multimodal AI, and cybersecurity into the open, giving developers, companies, and countries a great open foundation to build their own AI and defend themselves against cyber threats.

杰夫·布迪耶，Hugging Face 产品副总裁Jeff Boudier, VP Product, Hugging Face

强大的模型只是成功的一部分。开发者还需要自主决定模型的运行位置、优化方式，以及如何将其融入自己的系统。我们对 Mistral Large 4 感到振奋，因为开放权重让开发者能更好地掌控这些决策，而不必依赖单一的托管服务。在 ZML，我们认为，这种自主权对于将 AI 的进步转化为实用、可部署的软件至关重要。A powerful model is only part of the equation. Developers also need the freedom to decide where it runs, how it is optimized, and how it fits into their systems. We’re excited about Mistral Large 4 because open weights give builders more control over those decisions, rather than making them dependent on a single hosted service. At ZML, we see that freedom as essential to turning advances in AI into useful, deployable software.

ZML 首席执行官 Steeve MorinSteeve Morin, CEO of ZML

主权并不总是意味着必须自托管；有时，只需部署方式符合您的控制要求即可。对于倾向于使用托管服务的组织，Mistral Large 4 可通过 Microsoft Foundry 使用，并为 Azure 开发者提供私有部署选项。Sovereignty does not always require self-hosting; sometimes it just requires deployments that meet your control requirements. For organizations that prefer a managed service, Mistral Large 4 is available through Microsoft Foundry, with private deployment options for Azure developers.

开始使用 Le Chonk 构建Start building with Le Chonk

Mistral Large 4 现已开放体验、部署和定制。Mistral Large 4 is available to try, deploy, and customize today.

通过 Mistral Studio、Mistral API 或 OpenRouter 开始使用。若想获得完全掌控权，可从 Hugging Face 下载开放权重，并使用 vLLM 或 SGLang 部署到自己的基础设施上。模型卡和许可证包含了开始使用所需的一切信息；对于想深入了解的用户，完整的技术报告即将发布。Get started in Mistral Studio, through the Mistral API, or on OpenRouter. For full control, download the open weights from Hugging Face and deploy on your own infrastructure with vLLM or SGLang. The model card and license have everything you need to get started — and for those who want to go deeper, a full technical report is coming soon.

你也可以在 Vibe Code 中试用这个新模型：在你自己的代码库中，交给它一个功能来构建，或一个 bug 来修复。You can also try the new model in Vibe Code; give it a feature to build or a bug to fix in your own repository.

Mistral Large 4 也可通过 NVIDIA 托管的端点使用，它以经过优化的 NVIDIA NIM 容器形式提供支持，可通过 vLLM 部署服务。开发者可以使用开源 NeMo AutoModel 库进行定制，或使用开源 TensorRT LLM 库提供高吞吐量服务。Mistral Large 4, is also available through NVIDIA-hosted endpoints, as an optimized NVIDIA NIM container for serving with vLLM. Developers can use the open source NeMo AutoModel library for customization or the open source TensorRT LLM library for high-throughput serving.

企业还可以通过 Mistral Forge，针对专有数据、使用场景和防护规则定制 Mistral Large 4。欢迎联系我们的团队了解更多。Enterprises can also tailor Mistral Large 4 to their proprietary data, use cases, and guardrails with Mistral Forge. Talk to our team to learn more.

该模型可依据 Mistral Ecosystem Open-Weight 许可证使用，这是一种经过修改的 MIT 许可证。该模型可免费使用、修改、微调和商业化，但有一项条件：运营模型即服务（Model-as-a-Service）业务的企业，必须先从 Mistral 获得单独的商业许可证，才能将该模型用于商业用途。The model is available to use under the Mistral Ecosystem Open-Weight license, a modified MIT license. The model is free to use, modify, fine-tune, and commercialize, with a single condition: businesses that operate a Model-as-a-Service offering must obtain a separate commercial license from Mistral before using the model commercially.

从今天起，Mistral Large 4 的价格为：输入每百万 token 0.70 美元，缓存输入每百万 token 0.07 美元，输出每百万 token 2.10 美元。Starting today, Mistral Large 4 is priced at $0.70 input / M, $0.07 cached input / M, and $2.10 output / M tokens.

我们鼓励社区成员对 Mistral Large 4 进行测试。开展你自己的评估，挑战它的极限，并通过 X、Discord 或 GitHub 与我们分享你的发现。We encourage the community to put Mistral Large 4 to the test. Run your own evaluations, push its limits, and share your findings with us on X, Discord, or GitHub.

接下来呢？What's Next

Mistral Large 4 将成为新一代专用型和优化型 Mistral 模型的基础。在 Voxtral 和 Timestral 等近期发布成果的基础上，我们将继续拓展开放权重 AI 的前沿，为开发者和企业提供更大的自由，打造、定制并拥有自己的智能。Mistral Large 4 will serve as the foundation for a new generation of specialized and optimized Mistral models. Building on recent releases such as Voxtral and Timestral, we'll continue expanding the frontier of open-weight AI, continuing to give developers and enterprises greater freedom to build, customize, and own their intelligence.