# Thinking Machines Lab Inkling AI launch > Live situation record from CLSTR: https://clstr.news/situations/thinking-machines-lab-inkling-ai-launch > Updated: 2026-07-31T13:36:00.000Z. Sources: 41. Developments: 5. In mid‑July 2026, Thinking Machines Lab, the startup founded by former OpenAI CTO Mira Murati, released Inkling, a 975‑billion‑parameter multimodal mixture‑of‑experts model that activates roughly 41 billion parameters per query and supports a one‑million‑token context window. The launch was officially announced on 15 July 2026, and the model’s weights were made available under an Apache 2.0 licence on Hugging Face. The companion Tinker platform enables enterprises to fine‑tune Inkling on proprietary data for customized behaviour. Strategic infrastructure partners underpin the effort: Nvidia supplies gigawatt‑scale Vera Rubin compute capacity under a long‑term agreement, while Google Cloud has signed a multibillion‑dollar cloud‑services contract to support training and deployment. These collaborations are intended to position Inkling as an open‑weight, cost‑effective alternative to large proprietary models. Regional coverage in the Balkans highlighted the launch alongside Vodafone Albania’s pilot of a “Physical AI” system that autonomously repositions 4G and 5G antennas using AI‑driven sensors and robotic arms, illustrating a shift toward adaptive, efficient AI applications. Independent testing reported on 25 July 2026 showed Inkling achieving high benchmark scores (74.1 % on MCP Atlas and 77.6 % on SWE‑Bench) but struggling with complex coding, creative‑writing, and travel‑planning tasks. The model requires enterprise‑grade hardware and is priced at roughly $1 per million input tokens via OpenRouter, emphasizing privacy‑focused, compliance‑oriented use cases rather than everyday developer workloads. Later in July, the lab released Inkling Small, a 276‑billion‑parameter multimodal model with 12 billion active parameters. Weights were posted on Hugging Face on 30 July under an Apache 2.0 licence. The variant scores 40 on the Artificial Analysis Intelligence Index, outperforms the larger Inkling on several coding and reasoning benchmarks, and offers a 256 K‑token context window (noted as 1 million tokens in Japanese release notes). ## Claims - Inkling Small has 276 billion total parameters and 12 billion active parameters. (corroborated by 3 sources) - Inkling Small's weights are released under an Apache 2.0 license on Hugging Face. (corroborated by 3 sources) - Inkling Small scores 40 on the Artificial Analysis Intelligence Index, one point below Inkling's 41. (corroborated by 2 sources) - Inkling Small outperforms Inkling on Humanity's Last Exam (32 % vs. 30 %) and GPQA Diamond (89 % vs. 87 %). (corroborated by 2 sources) - Inkling Small averages 24 K output tokens per task, compared with 45 K for DeepSeek V4 Flash and 78 K for GPT‑5. (corroborated by 2 sources) - Inkling Small supports text, image, and speech inputs and has a 256 K-token context window. (corroborated by 2 sources) - Thinking Machines Lab released Inkling Small on July 30, 2026. (single source) - Inkling Small is about one‑quarter the size of Inkling (Inkling 975 B total, 41 B active). (single source) ## Timeline ### 2026-07-31: Thinking Machines Lab releases Inkling Small multimodal AI model Thinking Machines Lab released Inkling Small, a 276 B‑parameter multimodal AI model with open weights, Apache 2.0 license, and strong benchmark results, available on Hugging Face. 3 sources. https://clstr.news/cluster/thinking-machines-lab-releases-inkling-small-multimodal-ai-model ### 2026-07-25: Mira Murati's Inkling AI Model Tested Shows Mixed Performance Inkling, Mira Murati’s open‑source 975B‑parameter AI, scores high on benchmarks but falters in real‑world coding, writing and planning tasks, requiring heavy hardware and targeting compliance‑focused enterprise 2 sources. https://clstr.news/cluster/mira-muratis-inkling-ai-model-tested-shows-mixed-performance ### 2026-07-20: Balkan AI breakthroughs: Inkling model launch and Vodafone Albania pilot Inkling, a trillion‑parameter multimodal AI model, was launched by Thinking Machines Lab, while Vodafone Albania piloted autonomous AI‑driven antenna repositioning, highlighting advancing AI uses in the Balkans 3 sources. https://clstr.news/cluster/balkan-ai-breakthroughs-inkling-model-launch-and-vodafone-albania-pilot ### 2026-07-19: Thinking Machines Lab launches 975B-parameter Inkling AI model Thinking Machines Lab, led by Mira Murati, unveiled Inkling—a 975‑billion‑parameter open‑weight AI model with 41 billion active parameters per query, trained on 45 trillion multimodal tokens and backed by GPU‑c 3 sources. https://clstr.news/cluster/thinking-machines-lab-launches-975b-parameter-inkling-ai-model ### 2026-07-15: Thinking Machines Lab launches Inkling, 975‑billion‑parameter open‑weight AI model Thinking Machines Lab, founded by ex‑OpenAI CTO Mira Murati, unveiled Inkling – a 975‑billion‑parameter open‑weight multimodal model with 41 billion active parameters, available for free download and fine‑tuned 32 sources. https://clstr.news/cluster/thinking-machines-launches-inkling-an-openweight-ai-model --- Cite as: Thinking Machines Lab Inkling AI launch. CLSTR, https://clstr.news/situations/thinking-machines-lab-inkling-ai-launch