--- title: This $12 billion startup finally shipped something... description: Thinking Machines, the $12 billion startup founded by former OpenAI CTO Mira Murati, has released Inkling — a fully open-weights multimodal model with 970... canonical: https://daily.dev/posts/this-12-billion-startup-finally-shipped-something--p81y8qwfk twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: This $12 billion startup finally shipped something... | daily.dev og:description: Thinking Machines, the $12 billion startup founded by former OpenAI CTO Mira Murati, has released Inkling — a fully open-weights multimodal model with 970... og:url: https://daily.dev/posts/this-12-billion-startup-finally-shipped-something--p81y8qwfk og:image: https://api.daily.dev/og/posts/P81Y8QwFK.png og:image:alt: This $12 billion startup finally shipped something... og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # This $12 billion startup finally shipped something... **[Fireship](https://daily.dev/sources/fireship)** · 5 min read · 9 upvotes · 0 comments ## Summary Thinking Machines, the $12 billion startup founded by former OpenAI CTO Mira Murati, has released Inkling — a fully open-weights multimodal model with 970 billion total parameters (41B active per token via mixture-of-experts). Licensed under Apache and available on Hugging Face, it processes raw audio and pixels directly without separate encoders. While benchmarks place it mid-table among open models, its key differentiator is a 'thinking effort' dial for compute-cost tradeoffs and strong calibration via 'epistemics' training that makes it one of the best models at forecasting and admitting uncertainty. The business model pairs the free open model with Tinker, a paid fine-tuning API, letting users specialize Inkling for specific tasks without managing infrastructure. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### What is Inkling, the model released by Thinking Machines? Inkling is a fully open-weights mixture-of-experts model from Thinking Machines with 970 billion total parameters, though only about 41 billion activate per token via expert routing. It was pretrained on 45 trillion tokens of text, images, and audio, supports a 1 million token context window, and is released under an Apache license on Hugging Face. _Developers weighing open-weight model options for fine-tuning can follow releases like this one on daily.dev._ ### How does the thinking effort setting in Inkling work? Thinking effort is an adjustable dial that controls how much computation the model spends reasoning before answering. Turning it down produces cheap, fast responses, while turning it up lets Inkling match Kimi K2's performance on Terminal Bench while using roughly a third of the tokens, which matters significantly at high-volume agent scale. _Teams tuning cost versus reasoning quality in agent workloads can track model updates like this via daily.dev._ ### How many parameters does Kimi K2 from Moonshot have compared to other open models? Kimi K2 is a 2.8 trillion parameter model released by Moonshot that reportedly performs competitively with top-tier closed models like Gemini 3, and gained enough popularity to strain the company's compute capacity. It was announced the day after Thinking Machines released its own open-weights model, Inkling. _Anyone comparing large open-weight releases can keep up with fast-moving model launches through daily.dev._ ## Similar posts on daily.dev - [Inkling model from Thinking Machines Lab now on Databricks](https://daily.dev/posts/inkling-model-from-thinking-machines-lab-now-on-databricks-1wahujdlp) · databricks · 1 upvotes · 0 comments - [ThursdAI - Jul 16 - Inkling 975B open weights, Kimi K3 at 2.8T, a 27B model on a phone & Codex hits 9M](https://daily.dev/posts/thursdai---jul-16---inkling-975b-open-weights-kimi-k3-at-2-8t-a-27b-model-on-a-phone-codex-hits--hledq9pdh) · ThursdAI · 0 upvotes · 0 comments --- Tags: [#open-source](https://daily.dev/tags/open-source), [#llm](https://daily.dev/tags/llm), [#ai-agents](https://daily.dev/tags/ai-agents), [#deep-learning](https://daily.dev/tags/deep-learning), [#mixture-of-experts](https://daily.dev/tags/mixture-of-experts) [View this post on daily.dev](https://daily.dev/posts/this-12-billion-startup-finally-shipped-something--p81y8qwfk) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"This $12 billion startup finally shipped something...","url":"https://daily.dev/posts/this-12-billion-startup-finally-shipped-something--p81y8qwfk","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/this-12-billion-startup-finally-shipped-something--p81y8qwfk"},"datePublished":"2026-07-20T17:06:45.036Z","dateModified":"2026-09-13T21:18:48.347Z","description":"Thinking Machines, the $12 billion startup founded by former OpenAI CTO Mira Murati, has released Inkling — a fully open-weights multimodal model with 970...","image":"https://i.ytimg.com/vi/M51asSwRLxA/sddefault.jpg","thumbnailUrl":"https://i.ytimg.com/vi/M51asSwRLxA/sddefault.jpg","isAccessibleForFree":true,"articleSection":"Fireship","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"Fireship","logo":"https://media.daily.dev/image/upload/s--75ndqrkr--/f_auto,t_logo/v1702882094/logos/fireship.jpg","url":"https://daily.dev/sources/fireship"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/this-12-billion-startup-finally-shipped-something--p81y8qwfk","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":9},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"open-source,llm,ai-agents,deep-learning,mixture-of-experts","timeRequired":"PT5M","video":{"@type":"VideoObject","name":"This $12 billion startup finally shipped something...","description":"Thinking Machines, the $12 billion startup founded by former OpenAI CTO Mira Murati, has released Inkling — a fully open-weights multimodal model with 970...","thumbnailUrl":"https://i.ytimg.com/vi/M51asSwRLxA/sddefault.jpg","uploadDate":"2026-07-20T17:06:45.036Z","duration":"PT5M","url":"https://api.daily.dev/r/P81Y8QwFK","embedUrl":"https://www.youtube.com/embed/M51asSwRLxA"}} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Fireship","item":"https://daily.dev/sources/fireship"},{"@type":"ListItem","position":3,"name":"This $12 billion startup finally shipped something..."}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/this-12-billion-startup-finally-shipped-something--p81y8qwfk#faq","mainEntity":[{"@type":"Question","name":"What is Inkling, the model released by Thinking Machines?","acceptedAnswer":{"@type":"Answer","text":"Inkling is a fully open-weights mixture-of-experts model from Thinking Machines with 970 billion total parameters, though only about 41 billion activate per token via expert routing. It was pretrained on 45 trillion tokens of text, images, and audio, supports a 1 million token context window, and is released under an Apache license on Hugging Face. Developers weighing open-weight model options for fine-tuning can follow releases like this one on daily.dev."}},{"@type":"Question","name":"How does the thinking effort setting in Inkling work?","acceptedAnswer":{"@type":"Answer","text":"Thinking effort is an adjustable dial that controls how much computation the model spends reasoning before answering. Turning it down produces cheap, fast responses, while turning it up lets Inkling match Kimi K2's performance on Terminal Bench while using roughly a third of the tokens, which matters significantly at high-volume agent scale. Teams tuning cost versus reasoning quality in agent workloads can track model updates like this via daily.dev."}},{"@type":"Question","name":"How many parameters does Kimi K2 from Moonshot have compared to other open models?","acceptedAnswer":{"@type":"Answer","text":"Kimi K2 is a 2.8 trillion parameter model released by Moonshot that reportedly performs competitively with top-tier closed models like Gemini 3, and gained enough popularity to strain the company's compute capacity. It was announced the day after Thinking Machines released its own open-weights model, Inkling. Anyone comparing large open-weight releases can keep up with fast-moving model launches through daily.dev."}}]} ```