{"rewrite":{"id":"r_4294d29fe84d2c1cfbb7ffa7","clusterId":"c_57f4a047786a3b2fbd0c8e48","slug":"thinking-machines-lab-releases-inkling-an-open-weights-ai-model-for-fine-tuning","model":"deepseek-v4-flash","headline":"Thinking Machines Lab Releases Inkling, an Open-Weights AI Model for Fine-Tuning","summary":"Thinking Machines Lab has released Inkling, an open-weights AI model with 975 billion total parameters and 41 billion effective parameters using a Mixture-of-Experts architecture. The model handles text, images, audio, and video, supports up to 1 million tokens of context, and allows users to adjust the amount of compute used for reasoning. The company positions Inkling as a foundation for fine-tuning to specific business needs rather than a general-purpose model.","whyItMatters":"Inkling shifts the focus from raw benchmark performance to adaptability, offering a base model that organizations can fine-tune with their own data and rules.","webCardHtml":"\u003cp\u003eInkling was pre-trained on 45 trillion tokens covering text, images, audio, and video. The model uses a Mixture-of-Experts architecture with 975 billion total parameters but only 41 billion active per inference, keeping compute costs lower. Thinking Machines Lab designed Inkling as a base for fine-tuning rather than a finished product, arguing that organization-specific knowledge often matters more than general performance. Users can adjust the model\u0026#39;s \u0026#34;thinking effort\u0026#34; to trade speed for accuracy. On Terminal Bench 2.1, Inkling matched the performance of Nemotron 3 Ultra while generating about one-third the tokens. The full weights are available on Hugging Face, and a smaller variant called Inkling-Small is in testing.\u003c/p\u003e","blueskyPost":"Thinking Machines Lab released Inkling, an open-weights AI model with 975B total parameters and 41B active. It handles text, images, audio, video, and supports fine-tuning for specific business needs. The model's compute can be adjusted per task. Weights are on Hugging Face.","twitterPost":"Thinking Machines Lab released Inkling, an open-weights AI model with 975B total params (41B active). It handles text, images, audio, video, and is built for fine-tuning. Adjustable compute. Weights on Hugging Face. A smaller variant is coming.","threadsPost":null,"newsletterBlurb":"Thinking Machines Lab has released Inkling, an open-weights AI model designed for fine-tuning to specific business needs. The model uses a Mixture-of-Experts architecture with 975 billion total parameters and supports text, images, audio, and video. The company positions it as a foundation for organizations to adapt rather than a general-purpose model.","attributionJson":"[{\"source\":\"GIGAZINE\",\"url\":\"https://gigazine.net/news/20260716-inkling/\",\"title\":\"約1兆パラメーターで画像・音声の理解やコーディングに対応、用途別に調整できるオープンウェイトAIモデル「Inkling」登場\"}]","lintFlagsJson":null,"lintHits":0,"costUsd":0,"inputTokens":5402,"outputTokens":2195,"status":"published","repairAttempts":0,"nextRepairAt":null,"factsAttemptedAt":1784223102,"createdAt":"2026-07-16T17:23:16.000Z","publishedAt":"2026-07-16T17:27:28.000Z","updatedAt":"2026-07-16T17:23:16.000Z"},"cluster":{"id":"c_57f4a047786a3b2fbd0c8e48","canonicalTitle":"約1兆パラメーターで画像・音声の理解やコーディングに対応、用途別に調整できるオープンウェイトAIモデル「Inkling」登場","representativeArticleId":"a_f71c18e40ef805e85d506f52","sourceCount":1,"writtenSourceCount":1,"writeAttempts":0,"isSolo":true,"entitiesJson":"{\"anime_titles\":[],\"manga_titles\":[],\"work_titles\":[\"Inkling\"],\"studios\":[],\"people\":[],\"type\":\"announcement\",\"domain\":\"other\",\"is_roundup\":false}","contentType":"news","status":"published","firstSeenAt":"2026-07-16T02:15:00.000Z","lastSeenAt":"2026-07-16T02:15:00.000Z","updatedAt":"2026-07-16T17:27:29.000Z"},"attribution":[{"source":"GIGAZINE","url":"https://gigazine.net/news/20260716-inkling/","title":"約1兆パラメーターで画像・音声の理解やコーディングに対応、用途別に調整できるオープンウェイトAIモデル「Inkling」登場"}],"entities":{"anime_titles":[],"manga_titles":[],"work_titles":["Inkling"],"studios":[],"people":[],"type":"announcement","domain":"other","is_roundup":false},"keyFacts":null}