📊 Full opportunity report: The First Signs Of AI Change: What Thinking Machines’ Inkling Means on ThorstenMeyerAI.com — validation score, market gap, and execution plan.
TL;DR
Thinking Machines unveiled Inkling, a large open-weight AI model with 975 billion parameters, available on Hugging Face under Apache 2.0 license. This marks a shift toward more transparent, owner-controlled AI, though some restrictions may apply.
Thinking Machines has released its first foundation model, Inkling, openly available on Hugging Face under the Apache 2.0 license. This marks a significant shift in AI development, emphasizing model ownership and transparency over proprietary control.
Inkling is a Mixture-of-Experts transformer with 975 billion total parameters and 41 billion active, supporting a 1-million-token context window. It was pretrained on 45 trillion tokens across text, images, audio, and video, with a natively multimodal input design that processes text, images, and audio jointly without a vision adapter. The model was trained using a hybrid optimizer on NVIDIA systems, with over 30 million reinforcement learning rollouts improving reasoning performance.
The weights are released under Apache 2.0 license, allowing download, modification, and commercial use. However, reports indicate that Thinking Machines maintains a separate Model Acceptable Use Policy restricting surveillance, deception, and automated decision-making affecting individuals, which could limit the open-source nature of the model in practice. The full training data and pipeline are not publicly disclosed, a common industry norm but a point of contention for transparency advocates.
The weights came first: what Inkling actually signals
Mira Murati’s lab shipped its first foundation model — and the model isn’t the story. The order of operations is: full weights, Apache 2.0, day one, before any closed API. Plus a rare concession — the lab says it’s not the strongest model available, open or closed.
- AIME 2026 97.1%
- GPQA Diamond 87.2%
- MCP Atlas (Nemotron 44.7%) 74.1%
- VoiceBench · open-weight audio frontier 91.4%
- FORTRESS adversarial · best open 78.0%
- ForecastBench · calibration 61.1
- HLE text-only (GLM-5.2 40.1%) 29.7%
- SWE-bench Pro (GLM-5.2 62.1%) 54.3%
- Terminal-Bench 2.1 (GLM-5.2 82.7%) 63.8%
- SWE-bench Verified (Fable 5 95.0%) 77.6%
- Design Arena · 2nd open, behind GLM-5.2 ~10th
A 0.2 → 0.99 effort setting trades reasoning tokens against cost & latency, so you get a curve, not a point. On Terminal-Bench 2.1 it reportedly matches Nemotron 3 Ultra at ~⅓ the tokens. Peak score is a vanity metric when you serve millions of calls; the cost curve is what ships. (Bonus: its chain of thought compressed on its own during RL — nobody rewarded it; efficiency did.)
Pitched as the Western alternative to Chinese open weights (censorship-resistance training is the differentiator). But GLM-5.2 still wins on agentic/reasoning and Kimi K2.6 often on multimodal: best American open model, second in the open field. The irony — post-training was bootstrapped on synthetic data from Kimi K2.5.
BF16 needs ≥2 TB aggregate VRAM (8× B300 / 16× H200). NVFP4 still needs ≥600 GB. Not a workstation model — a 512 GB fleet falls just short. “Open” ≠ “runnable.” Mitigations: 1-bit GGUFs (~74% acc.), hosted eval routes, and Inkling-Small (12B active) — the release local-first builders actually want.
Open weights used to be a consolation prize. Inkling is a strategic open release — Apache 2.0, natively multimodal, honestly marketed, published complete on day one, optimized for deployment rather than headlines (the model isn’t the product; the fine-tuning platform is). It doesn’t need to win every benchmark for that to matter. The frontier is learning that owning the base beats renting the API — arriving now from the inside. For the sovereignty buyer: ① a real Western hedge against being switched off · ② verify the use policy before you build · ③ check the VRAM, then benchmark vs GLM-5.2 & Kimi K2.6 on your task.
Implications of Open-Weight AI Models for Industry
This release signifies a shift toward greater model ownership and transparency in AI development, allowing organizations to host and modify models independently. It challenges the traditional proprietary approach, potentially accelerating innovation and reducing reliance on closed APIs. However, the existence of a separate use policy raises questions about the true openness and enforceability of restrictions, making its long-term impact uncertain.
AI development model download
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Background of Open-Weight AI Model Releases
Until now, most large foundation models have been released as proprietary or with limited access, often via closed APIs. The recent trend has been toward commercial licensing, with some exceptions providing open weights but accompanied by restrictions. The release of Inkling under Apache 2.0 marks a notable departure, emphasizing model ownership and transparency. Historically, open models like GPT-2 and GPT-3 variants have influenced industry standards, but the scale and multimodal capabilities of Inkling represent a new frontier in open AI development.
“We believe in providing the community with powerful tools while maintaining responsible use policies.”
— Thinking Machines spokesperson
open-source AI models
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Questions About the Model’s Use Restrictions
It remains unclear how the separate Use Policy will be enforced and whether it will significantly limit the practical openness of Inkling. The policy reportedly prohibits surveillance, deception, and certain automated decisions, but the exact scope and enforceability are not publicly verified. Additionally, the full training data and pipeline are not disclosed, raising questions about transparency and reproducibility.
multimodal AI model
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Next Steps for Industry Adoption and Testing
Expect independent researchers and organizations to test and benchmark Inkling across various tasks, verifying claims and exploring its capabilities. Further details on the use policy and training data are anticipated, alongside potential updates or restrictions. Industry observers will monitor how the model’s open weights influence broader AI development and ownership models in the coming months.
large language model 975B parameters
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Key Questions
What makes Inkling different from other large AI models?
Inkling is notable for being openly available under the Apache 2.0 license, with full weights released publicly. It also features a multimodal, 975-billion-parameter architecture supporting a one-million-token context window, designed for flexible ownership and deployment.
Are there any restrictions on how I can use Inkling?
While the weights are openly licensed, reports suggest that Thinking Machines has a separate Acceptable Use Policy that restricts surveillance, deception, and certain automated decisions. The enforceability and scope of these restrictions are still unverified and require careful review before use.
Why is open licensing important for AI models?
Open licensing allows organizations to host, modify, and deploy models independently, reducing reliance on proprietary APIs. It fosters innovation, transparency, and potentially safer development by enabling community oversight and verification.
What are the potential risks of open-weight models?
Open weights can be misused for malicious purposes, such as generating disinformation or automating surveillance. Without clear, enforceable use restrictions, there is a risk of harm, especially if the model is used in sensitive domains.
What will happen next in the development of open AI models?
Researchers will benchmark Inkling’s performance, verify claims, and explore its capabilities. The industry will watch for updates on use policies, training data transparency, and how open-weight models influence AI ownership and safety standards.
Source: ThorstenMeyerAI.com
