📊 Full opportunity report: How An Inkling From Thinking Machines Could Change AI’s Course on ThorstenMeyerAI.com — validation score, market gap, and execution plan.
TL;DR
Thinking Machines has released Inkling, a large open-weight multimodal AI model, openly available on Hugging Face under Apache 2.0. This move challenges norms around model openness and ownership, with implications for AI development and regulation.
Thinking Machines has released its first foundation model, Inkling, openly available on Hugging Face under Apache 2.0 license. This marks a departure from typical industry practices, emphasizing open access over proprietary control, and could influence future AI development and ownership models.
The Inkling model is a Mixture-of-Experts transformer with 975 billion parameters, supporting a 1-million-token context window. It was trained on 45 trillion tokens of multimodal data, including text, images, audio, and video, with a native multimodal input design. The model’s weights are publicly available on Hugging Face under Apache 2.0, allowing users to download, modify, and deploy independently.
Unlike typical model releases, Thinking Machines explicitly stated that Inkling is not the strongest model available but prioritized transparency and open access. The company also indicated that a separate Model Acceptable Use Policy restricts certain applications, such as surveillance or deceptive practices, which complicates the notion of “open source.” The model was trained with a hybrid optimizer and involved over 30 million reinforcement learning rollouts, with some training data generated by open-weight models like Kimi K2.5.
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 Model Release for AI Ownership
The public release of Inkling under open licensing challenges the industry norm of proprietary models and raises questions about ownership, control, and regulation. By making the weights freely available, Thinking Machines enables organizations to fine-tune, inspect, and deploy the model independently, potentially accelerating innovation but also complicating oversight and safety measures.
This move could influence industry standards around transparency and open access, especially as concerns grow over AI safety, misuse, and accountability. However, the presence of a separate use policy layered on top of the open license introduces ambiguity about the scope of permissible uses, which could impact adoption and trust.
Accelerate Everything with Tensor Cores: A Developer’s Guide to High-Performance AI, Efficient Training, and Scalable Models
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Industry Norms and the Shift Toward Open Models
Traditionally, large foundation models are either proprietary or released with limited access, often through APIs. Recent efforts, like Meta’s Llama 2, have moved toward open weights, but with restrictions. Inkling’s release under Apache 2.0, combined with candid details about training and performance, marks a notable shift toward transparency and open ownership.
Thinking Machines, founded by former OpenAI CTO and staffed with team members involved in ChatGPT’s development, aims to challenge existing paradigms by prioritizing open access and honesty about model capabilities. The company’s decision to publish full weights immediately, rather than a closed API, underscores a different approach to AI deployment and ownership.
“Our goal is to empower developers and organizations with full access, while maintaining responsible use through our policies.”
— Thinking Machines spokesperson
ESP32 Basic Starter Ai Chatbot Kit Development Board USB-C Dual Core Microcontroller Support AP/STA/AP+STA Compatible with Arduino IDE IoT for Beginners Engineers
【High Performance】The ESP32 module is equipped with a dual-core CPU and features a Type-C USB interface, with 44…
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Unresolved Questions About Inkling’s Use and Control
It remains unclear how the separate Model Acceptable Use Policy will be enforced and whether it will effectively limit misuse. The distinction between open weights and restrictions layered on top raises questions about legal enforceability and practical control.
Additionally, the full scope of the training data and the potential for fine-tuning or repurposing by third parties are still to be examined. The impact of this release on industry standards and regulatory responses is also uncertain.
LLM INFERENCE ENGINEERING: Optimizing Large Language Models on NVIDIA GPUs
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Next Steps for Adoption and Industry Response
Expect independent researchers and organizations to test and benchmark Inkling’s capabilities further, especially in safety and bias. The company may also publish more details on the training data and use policy.
Regulators and industry groups could respond by reconsidering licensing norms and developing standards for open-weight models, especially regarding safety and misuse prevention. The model’s adoption and real-world impact will become clearer as it is integrated into applications.
AI-Powered UX Research: Run Research at the Speed Your Team Actually Needs
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 language models?
Inkling is a 975-billion-parameter multimodal model released openly under Apache 2.0, allowing independent use and modification, unlike most proprietary models.
Does open weights mean the model is entirely open source?
No. While the weights are openly available, the training data and full training pipeline are not published, and a separate use policy may impose restrictions.
What are the potential risks of releasing such a large open model?
Risks include misuse for harmful applications, bias amplification, and challenges in enforcing use policies. The layered restrictions add complexity to managing these risks.
How might this release influence the AI industry?
It could set a precedent for more open models and shift norms toward transparency, but also prompts debate over ownership, safety, and regulation.
Source: ThorstenMeyerAI.com
