📊 Full opportunity report: How Early Inkling From Thinking Machines Could Influence AI Development on ThorstenMeyerAI.com — validation score, market gap, and execution plan.
TL;DR
Thinking Machines has released Inkling, a large open-weight AI model, along with candid details on its training and licensing. This development raises questions about open-source practices and AI ethics, potentially influencing future AI strategies.
Thinking Machines has released its first foundation model, Inkling, under an open-source license, making its weights publicly available on Hugging Face. This marks a notable shift in AI development, emphasizing transparency and ownership, and directly challenges the industry norm of proprietary models.
Inkling is a 975-billion-parameter multimodal transformer supporting text, images, and audio inputs, with a 1-million-token context window. It was trained on 45 trillion tokens, including diverse media types, using a hybrid optimizer on NVIDIA systems. The model’s weights are available under Apache 2.0 license, allowing download, modification, and commercial use, representing a move toward more open AI ecosystems.
However, the announcement also notes a separate Model Acceptable Use Policy (AUP) that restricts certain uses, such as surveillance and deception, raising questions about the true openness of the release. The company emphasized that the weights are not open source in the traditional sense, as the training data and pipeline remain proprietary.
Early benchmark results show Inkling performing strongly in safety and multimodal tasks but only mid-tier in some language benchmarks. The release includes a smaller variant, Inkling-Small, which matches or surpasses the larger model on several tests. Full weights and further testing are expected after ongoing evaluation.
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 and Ethics
The release of Inkling under open licenses signals a potential shift toward more transparent and accessible AI development, enabling organizations to own and modify models independently. This move could challenge existing proprietary paradigms, fostering innovation and competition. However, the accompanying restrictions via the AUP highlight ongoing tensions around true openness, ethical use, and control, which could influence how future models are governed and adopted.
For AI developers, policymakers, and users, this development raises critical questions about data privacy, safety, and the limits of open-source licensing. The industry’s response to such models will shape the landscape of AI ownership, regulation, and ethical standards in the coming years.
Yahboom K230 AI Development Board 1.6GHz High-performance chip/2.4-inch Display/Open Source Robot Maker Python, Supports AI Visual Recognition CanMV Sensor (with Heightened Bracket)
- High-performance AI Chip: 1.6GHz processor with advanced AI capabilities
- Fast Response and High Load Handling: Supports real-time complex AI models
- Flexible Expansion Ports: Includes 12Pin GPIO for sensors and modules
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Background on Open-Weight AI Model Releases and Industry Norms
Until now, most large foundation models have been released with proprietary weights or limited access, emphasizing commercial control. Open-source initiatives, such as Meta’s Llama or EleutherAI’s models, have aimed for transparency but often with restrictions or incomplete data sharing. Thinking Machines’ approach with Inkling—offering full weights under Apache 2.0—represents a notable departure, aligning with broader efforts to democratize AI development.
The industry has recently faced debates over model safety, licensing, and control, especially after governments and organizations have taken steps to restrict or shut down certain models. The release of Inkling, despite its restrictions via the AUP, underscores a growing interest in balancing openness with responsible use, reflecting evolving norms in AI transparency and governance.
“Our goal was to provide a powerful, flexible model that organizations can own and tailor, while maintaining responsible use through our policies.”
— Thinking Machines spokesperson
Multimodal AI Systems Engineering: Building Production Vision-Language Models, Document AI, and Cross-Modal Retrieval Pipelines (Production AI Engineering Series)
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Unresolved Questions About Inkling’s Open Release and Use Policies
It remains unclear how strictly the Model Acceptable Use Policy will be enforced and how it might limit the practical openness of Inkling. The extent to which the proprietary training data and pipeline influence the model’s transparency is also uncertain. Additionally, the industry’s response to this hybrid approach—balancing open weights with restrictive policies—is still developing, and the long-term impact on AI innovation and regulation is unknown.
LOCAL LLM DEPLOYMENT: Training, Fine-Tuning, & Offline Inference: The Complete Developer’s Guide to Building, Training, and Running Private Open-Source AI Offline (with full source code)
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Next Steps in Testing, Policy Clarification, and Industry Adoption
Further independent testing of Inkling’s performance and safety benchmarks is expected as the model undergoes community evaluation. Clarifications on the scope and enforcement of the AUP are anticipated from Thinking Machines. Industry observers will monitor how competitors and regulators respond, potentially influencing future open-weight releases and licensing frameworks.
Introduction to AI Safety, Ethics, and Society
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Key Questions
Is Inkling truly open source?
No, while the weights are under Apache 2.0 license, the training data and pipeline are proprietary, and the company has a separate acceptable use policy that imposes restrictions.
What are the main restrictions on Inkling’s use?
The company’s policy reportedly prohibits surveillance, deception, and automated decision-making affecting individuals’ rights, but the enforceability and scope are still unclear.
How does Inkling compare to other models?
In benchmark tests, Inkling shows strong safety and multimodal capabilities but ranks mid-tier on some language benchmarks. Its open weights offer unique opportunities for customization and ownership.
What does this mean for AI regulation?
This development could influence future licensing and regulation strategies, emphasizing transparency and ownership but also highlighting ongoing debates over restrictions and true openness.
What is the significance of the open weights being under Apache 2.0?
The license allows free use, modification, and commercialization, but does not automatically ensure transparency of training data or enforce restrictions, which are governed separately.
Source: ThorstenMeyerAI.com
