1d:I[70616,["/_next/static/chunks/0kcf5dt6yxdox.js","/_next/static/chunks/1su5n1uv1ti4s.js","/_next/static/chunks/0gvke1wv6udfg.js","/_next/static/chunks/1dfrle3jgmlzh.js","/_next/static/chunks/162lgvduyr-fa.js","/_next/static/chunks/2gcnnopam8e9w.js","/_next/static/chunks/2g60iab5a1pkg.js"],"Breadcrumb"] 1e:I[90762,["/_next/static/chunks/0kcf5dt6yxdox.js","/_next/static/chunks/1su5n1uv1ti4s.js","/_next/static/chunks/0gvke1wv6udfg.js","/_next/static/chunks/1dfrle3jgmlzh.js","/_next/static/chunks/162lgvduyr-fa.js","/_next/static/chunks/2gcnnopam8e9w.js","/_next/static/chunks/2g60iab5a1pkg.js"],"ArticleGrid"] 1f:T1788,

Inception Labs is making an unusual pitch with Mercury 2.5: not a smarter model, but one fast and cheap enough to be called dozens of times inside a single user interaction. The startup says it is the largest diffusion language model ever trained, clocks 1,107 tokens per second on widely available NVIDIA GPUs, and lists at $0.20 and $0.75 per million input and output tokens — discounted 80%, to $0.04 and $0.15, at launch.

Key takeaways

Mercury 2.5 delivers a claimed 40% intelligence gain over Mercury 2 while keeping the same low-latency, low-cost serving profile.
Inception positions its quality against cost-optimised frontier tiers — GPT-5.6 Luna (Low), Gemini 3.5 Flash-Lite and Claude Haiku 4.5 — rather than against flagship reasoning models.
Augment Code reports that moving context compaction to Mercury cut latency 82%, from roughly 150 seconds to 27 seconds, and reduced cost by 90%.

Why a diffusion LLM competes on latency

Conventional large language models emit one token after another, so wall-clock time scales with how much they say. Diffusion models refine a whole block of text across parallel passes instead, and that architectural difference is why Inception Labs can quote four-digit throughput on commodity accelerators rather than on specialised inference silicon.

The company frames Mercury 2.5 as the first output of a feedback loop rather than a benchmark push: usage grew more than an order of magnitude after Mercury 2 shipped, across thousands of developers and dozens of enterprise deployments, and Inception says it rebuilt its evaluations around the production failure cases that growth exposed before starting training.

Where customers are running it

Search shows why per-call latency compounds. One query can fan out into dozens of model invocations — planning, query rewriting, reranking, fact structuring, summarisation, answer checking — and the whole chain has to finish before the user notices a pause, which is why Inception says several leading search-infrastructure companies now run Mercury in production.

Voice agents are the harder constraint, because latency there is audible as dead air on a phone call. OpenCall, which builds AI phone agents for live customer conversations, measures median model response latency near 170 milliseconds on its production workload, and co-founder and CEO Oliver Silverstein says the switch cut the company's P99 from several minutes to one second and its P50 from 0.4 seconds to under 0.2:

significantly faster than any other provider we've seen, and that's including reasoning.

Coding agents present the third pattern, and the numbers there are the most concrete in the announcement. Augment Code moved context compaction onto Mercury and watched it drop from roughly 150 seconds to 27 — an 82% cut — while costing 90% less at what the company says is unchanged quality, with MCP tool-search summaries returning in under a second.

Two previews aimed at the latency floor

Mercury Voice, previewed alongside the main release, is a dLLM tuned for the tightest budgets in spoken interaction, with time-to-first-token under 170 milliseconds. A second preview, Mercury Router, uses a diffusion model to read an incoming prompt and pick a downstream model, open or closed, on quality-speed-cost grounds.

Shruti Koparkar, a senior product manager in NVIDIA's Accelerated Computing Group, read the release as a marker of how quickly a new architecture can harden into production-ready systems on existing AI infrastructure. That framing is the real claim here: diffusion for text has spent years as a research curiosity, and speed numbers alone have not moved buyers before — OpenAI's own ultrafast tier leaned on Cerebras silicon to get a comparable jump, where Inception is claiming the gain comes from the architecture rather than the hardware.

Availability and what comes next

Mercury 2.5 serves through the Inception API, Baseten and OpenRouter, with 100 million free tokens for new API users and enterprise options covering dedicated capacity, autoscaling, compliance controls and configurable data retention. Tunable reasoning, parallel tool calls and schema-aligned JSON round out the feature set, the last mattering for agent pipelines that depend on structured output.

A successor is already in training. Inception describes it as its largest model yet, targeted for release within months, and says the goal is a capability leap that does not surrender diffusion's speed or token efficiency.

FAQ

How much does Mercury 2.5 cost?

List pricing is $0.20 per million input tokens and $0.75 per million output tokens. At launch Inception is discounting that by 80%, to $0.04 per million input and $0.15 per million output. New API users also get 100 million free tokens.

Is Mercury 2.5 open source?

No. Mercury 2.5 is a proprietary hosted model served through the Inception API, Baseten and OpenRouter. Enterprise customers can get dedicated capacity and configurable data retention, but the weights are not published.

What is a diffusion LLM?

A diffusion language model generates text by iteratively refining a block of tokens in parallel rather than predicting them strictly left to right. The approach trades architectural familiarity for throughput, which is why it is aimed at latency-sensitive work such as voice agents, search pipelines and agent scaffolding.

a:["$","div",null,{"className":"mx-auto max-w-7xl px-4 py-8 sm:px-6","children":[["$","script",null,{"type":"application/ld+json","dangerouslySetInnerHTML":{"__html":"{\"@context\":\"https://schema.org\",\"@type\":\"BreadcrumbList\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https://ainewsway.com\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"#voice-ai\",\"item\":\"https://ainewsway.com/tag/voice-ai\"}]}"}}],["$","script",null,{"type":"application/ld+json","dangerouslySetInnerHTML":{"__html":"{\"@context\":\"https://schema.org\",\"@type\":\"ItemList\",\"@id\":\"https://ainewsway.com/tag/voice-ai#itemlist\",\"name\":\"#voice-ai Articles\",\"numberOfItems\":1,\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"url\":\"https://ainewsway.com/articles/mercury-25-ships-at-1107-tokens-a-second-and-four-cents-a-million-mu10wirs\",\"name\":\"Mercury 2.5 Ships at 1,107 Tokens a Second and Four Cents a Million\"}]}"}}],["$","$L1d",null,{"items":[{"label":"# voice-ai"}]}],["$","h1",null,{"className":"mb-8 text-3xl font-bold text-zinc-900 dark:text-zinc-100","children":["#","voice-ai"]}],[["$","$L1e",null,{"articles":[{"id":"3ac02a00-a234-494a-8b79-36c4190c2852","slug":"mercury-25-ships-at-1107-tokens-a-second-and-four-cents-a-million-mu10wirs","title":"Mercury 2.5 Ships at 1,107 Tokens a Second and Four Cents a Million","subtitle":"Inception Labs is selling throughput, not intelligence — and customers in voice and coding agents are buying it","content":"$1f","excerpt":"A phone agent startup says Mercury cut its worst-case response from minutes to one second. Inception Labs' new diffusion model is built around that kind of number.","coverImage":"https://cdn.ainewsway.com/uploads/2026/09/32f93c7f-9ef6-45fc-9948-aede0461058a.jpg","coverImageCaption":"GPU racks in a research data centre — Inception Labs says Mercury 2.5 reaches 1,107 tokens per second on widely available NVIDIA hardware rather than specialised inference silicon.","authorId":"6773f67d-d7fa-4b19-8f38-384f10ac761b","status":"PUBLISHED","publishedAt":"2026-09-14T09:14:01.235Z","viewCount":0,"likeCount":0,"readingTime":5,"isFeatured":false,"isBreakingNews":false,"metaTitle":"Mercury 2.5: 1,107 Tokens a Second, $0.04 a Million","metaDescription":"Inception Labs' Mercury 2.5 claims the largest diffusion LLM ever trained at 1,107 tokens/sec and $0.20 per million input tokens. Pricing and customers.","metaKeywords":["Mercury 2.5","Inception Labs","diffusion LLM","dLLM","LLM latency","AI inference cost","voice agents","Mercury Router"],"ogImage":null,"aiSummary":"Inception Labs released Mercury 2.5, which it describes as the largest diffusion language model ever trained, running at 1,107 tokens per second on widely available NVIDIA GPUs. It lists at $0.20 per million input and $0.75 per million output tokens, discounted 80% at launch, with a claimed 40% intelligence gain over Mercury 2. The pitch is throughput for agent pipelines: Augment Code cut context compaction from 150 seconds to 27 and costs by 90%.","aiTags":[],"sentiment":null,"articleType":"NEWS","videoEmbed":null,"videoType":null,"seoScore":85,"isAutoGenerated":true,"isBackfill":false,"pipelineStage":null,"copyrightScore":null,"similarityScore":null,"matchedSourceId":null,"createdAt":"2026-09-14T09:11:03.690Z","updatedAt":"2026-09-14T12:10:39.059Z","contentUpdatedAt":"2026-09-14T12:10:39.058Z","deletedAt":null,"author":{"id":"6773f67d-d7fa-4b19-8f38-384f10ac761b","name":"Seung Jung","username":"Seung Jung","avatar":null},"categories":[{"id":"b84a09d8-4faa-4fea-8c60-a37c2cf9f593","name":"LLM & Chatbots","slug":"llm-chatbot"}],"tags":[{"id":"968da395-163c-463e-a995-075699e5d7a4","name":"Diffusion LLM","slug":"diffusion-llm"},{"id":"e909dcd8-f32f-49b1-9b71-2f5bc89d99ee","name":"LLM Pricing","slug":"llm-pricing"},{"id":"661a50e8-d3f4-456c-9773-9ffc9be024e7","name":"AI Inference","slug":"ai-inference"},{"id":"9e33fca4-39a1-4abe-b37b-4480c20aa54d","name":"Voice AI","slug":"voice-ai"},{"id":"51075b6e-7339-4ca5-a00d-1b068653c310","name":"NVIDIA","slug":"nvidia"},{"id":"18cdd61e-c75a-4aff-843e-db288dddc418","name":"Coding Agents","slug":"coding-agents"}],"translations":[{"id":"03ea6646-cee4-416e-b80c-20212b9509c8","language":"ko","title":"Mercury 2.5, 초당 1,107토큰에 100만 토큰당 4센트로 나왔다","subtitle":"Inception Labs는 지능이 아니라 처리 속도를 판다. 음성·코딩 에이전트 고객이 여기에 지갑을 열었다","excerpt":"전화 에이전트 스타트업은 Mercury가 최악의 응답 시간을 수 분에서 1초로 줄였다고 밝혔다. Inception Labs의 새 디퓨전 모델은 그런 숫자를 겨냥해 만들어졌다.","metaTitle":"Mercury 2.5: 초당 1,107토큰, 100만 토큰당 4센트","metaDescription":"Inception Labs의 Mercury 2.5는 역대 최대 규모 디퓨전 LLM으로 초당 1,107토큰, 입력 100만 토큰당 0.20달러를 내세운다. 가격과 도입 사례를 정리했다."},{"id":"a01f9c57-d5f4-472f-b9e4-c952d91f69fd","language":"ja","title":"Mercury 2.5、毎秒1,107トークンで100万トークン4セント","subtitle":"Inception Labsが売るのは知能ではなくスループットで、音声エージェントやコーディングエージェントの顧客がそれを買っています","excerpt":"電話エージェントのスタートアップは、Mercuryによって最悪時の応答が数分から1秒に縮んだと述べています。Inception Labsの新しい拡散モデルは、その種の数字を軸に設計されています。","metaTitle":"Mercury 2.5: 毎秒1,107トークン、100万で0.04ドル","metaDescription":"Inception LabsのMercury 2.5は史上最大の拡散LLMとして毎秒1,107トークン、入力100万トークン0.20ドルを掲げます。価格と導入事例をまとめました。"},{"id":"f620a747-6e64-4eb0-ae2e-9f45bb34cb5d","language":"zh","title":"Mercury 2.5 上线：每秒1,107个词元，百万词元四美分","subtitle":"Inception Labs 卖的不是智能而是吞吐量，语音与编程智能体的客户正在买单","excerpt":"一家电话智能体初创公司称，Mercury 把最差情况下的响应时间从数分钟压到了一秒。Inception Labs 的新扩散模型正是围绕这类数字打造的。","metaTitle":"Mercury 2.5：每秒1,107词元，百万词元0.04美元","metaDescription":"Inception Labs 的 Mercury 2.5 号称史上最大扩散语言模型，每秒1,107词元，输入百万词元0.20美元。本文梳理定价与落地客户。"},{"id":"d4cae5fb-4209-4fec-bffb-45f3841a7977","language":"es","title":"Mercury 2.5 llega a 1.107 tokens por segundo y cuatro centavos el millón","subtitle":"Inception Labs vende rendimiento, no inteligencia, y los clientes de agentes de voz y de código están comprando","excerpt":"Una startup de agentes telefónicos afirma que Mercury recortó su peor tiempo de respuesta de minutos a un segundo. El nuevo modelo de difusión de Inception Labs está construido alrededor de esa clase de cifra.","metaTitle":"Mercury 2.5: 1.107 tokens por segundo y USD 0,04 el millón","metaDescription":"Mercury 2.5 de Inception Labs dice ser el mayor LLM de difusión entrenado: 1.107 tokens por segundo y USD 0,20 por millón de tokens de entrada."}],"celebrities":[{"id":"b826d323-56d8-4894-9c87-138b9e0c0f88","nameKo":"인셉션 랩스","nameEn":"Inception Labs","slug":"inception-labs","profileImage":null},{"id":"95f8f574-a1e4-4a77-8783-1419c1421b78","nameKo":"엔비디아","nameEn":"NVIDIA","slug":"nvidia","profileImage":null}]}]}],"$L20"]]}]

#voice-ai

Mercury 2.5 llega a 1.107 tokens por segundo y cuatro centavos el millón

Suscríbete al Boletín

#voice-ai

Mercury 2.5 llega a 1.107 tokens por segundo y cuatro centavos el millón