# Inception Labs 发布扩散大语言模型 Mercury 3,速度达 1068 tokens/秒
> 原标题:Diffusion at Scale: Introducing Mercury 3
- 来源:Inception Labs:Blog
- 发布时间:2026-10-19T07:00:00.000Z
- AIHOT:https://aihot.news/items/osk3jj5jltogwr6zpo09w80j8
- 原文:https://www.inceptionlabs.ai/blog/introducing-mercury-3
## 摘要
Inception Labs 发布扩散式大语言模型 Mercury 3,称其在多项基准上超过 GPT-6 Luna 和 Claude Haiku 5.5,速度达 1068 tokens/秒,参数量较 Mercury 1 扩大 40 倍以上。
## 推荐理由
官方发布给出扩散模型规模化的关键数据,包括速度、上下文和价格,可用作评估实时 AI 应用选型的参考。
## 正文 · 中文译文
今天我们推出 Mercury 3,这是迄今为止发布的最强大的基于扩散的 LLM。在一系列基准测试中,Mercury 3 的表现优于 GPT-6 Luna 和 Claude Haiku 5.5,同时运行速度快 4–8 倍,达到每秒 1,000+ 个 token。
Mercury 3 证明了文本扩散可以扩展。它的规模比 Mercury 1 大 40 倍以上,智能程度也大幅提升,但速度同样快。

早期访问合作伙伴已经在使用 Mercury 3 进行语音和聊天支持、实时辅导、数据工程和网页导航:
## Mercury 3 一览
- **质量:**与 GPT-6 Luna 和 Claude Haiku 5.5 等前沿速度优化模型持平或更优。
- **速度:**在 NVIDIA GPU 上每秒 1,068 个 token。
- **上下文:**1M token,高于 Mercury 2.5 的 260K
- **价格:**每百万输入 token 0.20 美元,每百万输出 token 0.70 美元。
- 发布时,Mercury 3 享受 50% 折扣,每 1M 输入 token 0.10 美元,每 1M 输出 token 0.35 美元。缓存 token 在输入价格基础上享受 90% 折扣。

## 大规模扩散
当我们推出 Mercury 1 时,它是扩散 LLM 可以为商业应用构建的第一个证明点。AI 社区悬而未决的问题是,扩散 LLM 能否扩展到在复杂用例上与自回归模型竞争。
Mercury 3 回答了这个问题:它们可以。我们将 Mercury 的总参数量扩展了 40 倍以上,在一系列标准基准上的智能水平从 Mercury 1(mini)到 Mercury 3 提升了 3 倍。推动自回归模型进步的缩放定律同样适用于扩散。

## 生产环境中的 Mercury 3
早期访问合作伙伴已经在运行 Mercury 3,为浏览器使用智能体、互动教育、实时搜索智能体和客户支持等对延迟敏感的应用提供支持。
### 浏览器使用智能体
智能体在每一步都要等待模型,因此小的延迟会迅速累积。无论智能体是在浏览网站、跨来源研究,还是分析日志和追踪事件,Mercury 都能让每一轮保持快速。智能体能更早完成,并在回答前有时间尝试更多方法。
我们在生产环境的网页智能体框架中,针对前沿模型,在 50 个实时 Mind2Web 导航任务上运行了 Mercury 3。Mercury 3 完成了其中 50%,显著多于 GPT-6.1-Sol。它使用与前沿模型相同的 token 量和相同的智能体步数做到这一点,而每 token 成本仅为其一小部分。
Shuhao Zhang,TinyFish AI 联合创始人
### 互动教育
学习依赖快速反馈循环,任何延迟都会让学生脱离材料。Mercury 对每个请求进行推理并近乎实时地响应,无论是进行对话的语音辅导老师、批改作业的助手,还是根据学生刚做错的内容调整下一道题的应用。
Mercury 3 让我们能够为每个学生提供即时反馈,同时不放弃推理。提示、批改答案和定义现在大约一秒内返回,质量与我们之前运行的模型一样好或更好。当学生卡在一道题上时,这种速度让他们继续思考而不是等待,而且它可以无缝替换我们现有的技术栈。
Michael Tong,General Learning(又名 RevisionDojo)联合创始人/CTO
### 实时搜索智能体
搜索应用在仅仅几秒延迟后就开始让人感觉无响应。Mercury 3 运行速度足够快,可以处理数十份文档,并在用户仍保持参与时给出答案。
在智能眼镜上,一个需要五秒才能给出的答案会让人觉得系统坏了。在我们的 Sense API 测试中,Octen 搜索耗时约 80ms;模型生成才是瓶颈。在我们测试过的快速模型中,Mercury 3 在速度与质量之间取得了最佳平衡,生成时间中位数为 0.9 秒,端到端响应时间中位数为 2.2 秒。
Kuan Zou,Octen AI 创始人兼 CEO
### 客户支持
支持人员必须快速作答,而且往往服务于受监管行业,答错的代价很高。Mercury 让回复保持对话感,同时不牺牲准确性。
在 Lorikeet,我们帮助复杂且受监管的企业为客户构建礼宾服务,覆盖语音、聊天和电子邮件。在聊天和语音场景中,客户需要快速响应,但我们不能在准确性或正确完成任务上妥协。在早期测试中,Mercury 3 在延迟上优于其他模型,同时在评估和生产环境中的表现也更高。
Michael Gribben,Lorikeet 技术团队成员
## 实时产品组合
Mercury 3 是不断扩大的实时 AI 产品组合中的旗舰模型。
- **Mercury Voice:**Mercury Voice 针对语音应用极低的延迟要求进行了定制,并致力于最小化首个答案 token 的时间(TTFAT),而不仅仅是端到端延迟。它已在 Audivi AI、Altur 和 OpenCall 为智能体提供支持。
- **Mercury Decide:**Mercury Decide 是专为大批量分类构建的决策模型。它在包括 JevBench 和 LegalBench 在内的基准测试中优于 Jev。
## 开始使用
Mercury 3 现已在 Inception API Platform 上线,并可通过 OpenRouter、Baseten 和 Vercel 使用。它兼容 OpenAI API,因此无需重写即可接入你现有的技术栈。企业部署支持专用容量、自动扩缩容、合规控制和可配置的数据保留。
- [**在聊天中试用 Mercury 3**](https://chat.inceptionlabs.ai/)
- [**试用 API**](https://platform.inceptionlabs.ai/),获赠 1 亿免费 token
- [**阅读 API 文档**](https://docs.inceptionlabs.ai/get-started/get-started)
- **企业评估:**我们将与你合作,围绕工作负载适配、评估设计以及在你服务约束下的性能展开工作。[**联系我们**。](https://www.inceptionlabs.ai/enterprise#contact-sales)
## Mercury 3 已上线。扩散模型正在扩展。
## 正文 · 原文
Today we’re introducing Mercury 3, the most capable diffusion-based LLM released to date. Across a range of benchmarks, Mercury 3 outperforms GPT-6 Luna and Claude Haiku 5.5, while running 4–8x faster at 1,000+ tokens per second.
Mercury 3 demonstrates that text diffusion scales. It’s more than 40x larger than Mercury 1 and is vastly more intelligent, but just as fast.

Early-access partners already use Mercury 3 for voice and chat support, real-time tutoring, data engineering, and web navigation:
## Mercury 3 at a glance
- **Quality:** On par or better than frontier speed-optimized models like GPT-6 Luna and Claude Haiku 5.5.
- **Speed:** 1,068 tokens per second on NVIDIA GPUs.
- **Context:** 1M tokens, up from 260K for Mercury 2.5
- **Price:** $0.20 per million input and $0.70 per million output tokens.
- At launch, Mercury 3 is 50% off at $0.10 per 1M input and $0.35 per 1M output. Cached tokens are 90% off the input price.

## Diffusion at scale
When we launched Mercury 1, it was the first proof point that diffusion LLMs could be built for commercial applications. The open question in the AI community was whether diffusion LLMs could scale to compete with autoregressive models on complex use cases.
Mercury 3 answers that question: they can. We’ve scaled Mercury more than 40x in total parameters, and intelligence across a range of standard benchmarks has risen 3x from Mercury 1 (mini) to Mercury 3. The scaling laws that drove progress in autoregressive models hold for diffusion too.

## Mercury 3 in production
Early-access partners are already running Mercury 3 to power latency-sensitive applications such as browser use agents, interactive education, real-time search agents, and customer support.
### Browser Use Agents
Agents wait on the model at every step, so small delays compound quickly. Mercury keeps each pass fast, whether the agent is navigating websites, researching across sources, or analyzing logs and tracing incidents. Agents finish sooner and have time to try more approaches before answering.
We ran Mercury 3 inside our production web-agent harness against frontier models, on 50 live Mind2Web navigation tasks. Mercury 3 completed 50% of them, significantly more than GPT-6.1-Sol. It did so using the same token volume and the same number of agent steps as a frontier model, at a fraction of the per-token cost.
Shuhao Zhang, Cofounder, TinyFish AI
### Interactive Education
Learning runs on fast feedback loops, and any delay pulls a student out of the material. Mercury reasons through each request and responds in near real time, whether it’s a voice tutor carrying a conversation, an assistant grading work, or an app adapting the next question to what the student just got wrong.
Mercury 3 allowed us to give every student instant feedback without giving up reasoning. Hints, marked answers and definitions now come back in about a second, with quality as good as or better than the models we ran before. When a student is stuck on a question, that speed keeps them working instead of waiting, and it was a drop-in swap for our existing stack.
Michael Tong, Co-founder/CTO, General Learning aka RevisionDojo
### Real-time search agents
Search applications start to feel unresponsive after just a few seconds delay. Mercury 3 runs fast enough to process dozens of documents and still answer while the user is engaged.
On smart glasses, an answer that takes five seconds feels broken. In our Sense API tests, Octen search took about 80ms; model generation was the bottleneck. Mercury 3 delivered the best balance of speed and quality among the fast models we tested, with a median generation time of 0.9 seconds and a median end-to-end response time of 2.2 seconds.
Kuan Zou, Founder & CEO, Octen AI
### Customer support
Support agents have to answer quickly, often for regulated businesses where a wrong answer is costly. Mercury keeps responses conversational without sacrificing accuracy.
At Lorikeet, we help complex and regulated businesses build a concierge for their customers that works across voice, chat and email. In chat and voice, customers need fast responses, but we can't compromise on accuracy or getting the job done correctly. In early testing, Mercury 3 beat other models on latency while also performing higher on evals and in production.
Michael Gribben, Member of Technical Staff, Lorikeet
## A real-time product portfolio
Mercury 3 is the flagship model in a widening portfolio of products that enable real-time AI.
- **Mercury Voice:** Mercury Voice is customized for the extreme low-latency requirements of voice applications and for minimizing time to first answer token (TTFAT), not just end-to-end latency. It’s already powering agents at Audivi AI, Altur, and OpenCall.
- **Mercury Decide:** Mercury Decide is a decision model built for high-volume classification. It outperforms Jev on benchmarks including JevBench and LegalBench.
## Get started
Mercury 3 is available today on the Inception API Platform and through OpenRouter, Baseten, and Vercel. It is OpenAI API compatible, so it drops into your existing stack with no rewrites. Enterprise deployments support dedicated capacity, autoscaling, compliance controls, and configurable data retention.
- [**Try Mercury 3 in chat**](https://chat.inceptionlabs.ai/)
- [**Try the API**](https://platform.inceptionlabs.ai/) with 100 million free tokens
- [**Read the API docs**](https://docs.inceptionlabs.ai/get-started/get-started)
- **Enterprise evaluations:** we'll partner with you on workload fit, eval design, and performance under your serving constraints. [**Contact us**.](https://www.inceptionlabs.ai/enterprise#contact-sales)
## Mercury 3 is live. Diffusion scales.