📊 Full opportunity report: The Significance Of Weights And Inkling In AI’s Future on ThorstenMeyerAI.com — validation score, market gap, and execution plan.
TL;DR
Thinking Machines released its Inkling model with open weights on Hugging Face, emphasizing transparency and openness. The model is not the strongest available but represents a key step toward open AI development. Uncertainties remain about licensing details and future testing results.
Thinking Machines has released the full weights of its Inkling model under an open license, making it accessible for download and modification. This move is significant because it challenges the industry norm of proprietary models and emphasizes transparency in AI development, impacting how organizations may adopt and build upon large foundation models.
The Inkling model 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, and is natively multimodal with input types including text, images, and audio. The model is available on Hugging Face under Apache 2.0 license, allowing free download, modification, and commercial use, which offers a notable alternative to proprietary models.
Thinking Machines also announced a smaller variant, Inkling-Small, with 276 billion total parameters and 12 billion active, which has shown promising benchmark performance. The training process involved hybrid optimization and over 30 million reinforcement learning rollouts, with some training data generated from open-weight models like Kimi K2.5, a Chinese model, highlighting a layered approach to model training.
However, the release has nuances: the weights are not open source, and the full training data and pipeline remain proprietary. Additionally, reports suggest that Thinking Machines maintains a separate Model Acceptable Use Policy (AUP), which could impose restrictions on use, raising questions about the true openness of the model despite its Apache 2.0 license.
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 Weights for AI Development
The release of Inkling’s full weights under an open license represents a potential shift toward greater transparency and democratization in AI. It allows organizations and researchers to inspect, modify, and deploy the model independently, reducing reliance on closed APIs and fostering innovation. This move could influence industry standards, encouraging more companies to release open weights and challenge proprietary dominance.
At the same time, the existence of a separate Acceptable Use Policy (AUP) layered on top of the open license introduces ambiguity. If restrictions are enforced, they could limit the model’s openness in practice, especially in sensitive domains like surveillance or automated decision-making. The development raises important questions about what “open” truly means in AI licensing and how such layered policies will be managed.
HP 15.6" Laptop – Complete Productivity Solution, Windows 11, Microsoft Office, Intel 4 Core N100, 16GB RAM, 640GB Storage (128GB UFS + 512GB SD Card), Copilot AI, Lightweight – Silver
Upgraded to 16GB High-Speed Memory Boosted to 16GB RAM for improved multitasking — switch between apps, browser tabs,…
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Industry Trends Toward Transparency and Openness
Over recent years, the AI community has seen a growing push for open models, driven by concerns over transparency, safety, and democratization. Notable examples include Meta’s Llama model and Meta’s release of Llama 2 under open licenses, sparking debate about responsible openness. Historically, many large models have been kept proprietary due to commercial and safety considerations.
Thinking Machines’ decision to release Inkling’s weights openly, despite not being the strongest model, signals a strategic emphasis on transparency. The company’s approach contrasts with other industry players who prioritize control over access, and it aligns with broader movements advocating for open science and community-driven AI development.
“We believe in providing the community with the tools to innovate freely, which is why we released Inkling’s weights openly, even if it isn’t the top-performing model.”
— Thinking Machines spokesperson
HIWONDER Humanoid Robot with ChatGPT Multimodal AI Models AI Embodied Intelligent Vision Scene Voice Understanding 20DOF Educational Robotic Kit Python Programming, TonyPi Pro & RaspberryPi 5 16GB
Al-Driven & Raspberry Pi Powered. TonyPi is a high-performance AI vision robot designed for AI education applications. It…
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Unresolved Questions About Licensing and Use Restrictions
It remains unclear how the separate Model Acceptable Use Policy (AUP) will be enforced and whether it will impose restrictions that effectively limit the model’s openness. The exact scope of these restrictions, their legal enforceability, and how they might impact commercial or research use are still under question. Additionally, the full training data and pipeline have not been disclosed, leaving gaps in understanding the model’s provenance and safety measures.
Ollama – Open Source AI Tool Complete Guidebook
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Next Steps in Model Testing and Policy Clarification
Further independent testing and benchmarking of Inkling and Inkling-Small are expected to evaluate their performance across diverse tasks. Clarification from Thinking Machines on the AUP and licensing details will be crucial for organizations considering adoption. The community will likely monitor how the model performs in real-world applications and whether the layered restrictions are enforced.
Additionally, other organizations may follow suit, releasing open weights with layered policies, which could reshape norms around transparency, safety, and commercial use in AI development.
NVD RTX PRO 6000 Blackwell Professional Workstation Edition Graphics Card for AI, Design, Simulation, Engineering – 96GB DDR7 ECC Memory – 4th Gen RT/5th Gen Tensor Core GPU – OEM Packaging
[NVIDIA Blackwell Streaming Multiprocessor] The new SM features increased processing throughput, and new neural shaders that integrate neural…
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Key Questions
What makes Inkling different from other foundation models?
Inkling is a Mixture-of-Experts transformer with 975 billion parameters, supporting multimodal inputs and a 1-million-token context window. It is notable for being released with open weights under Apache 2.0, allowing free use and modification, unlike many proprietary models.
Are the weights of Inkling truly open source?
The weights are released under Apache 2.0 license, which permits use, modification, and commercialization. However, the full training data, pipeline, and any layered usage restrictions via the AUP mean it is not fully open source in the traditional sense.
Why does the layered Acceptable Use Policy matter?
If the AUP imposes restrictions on surveillance, deception, or automated decision-making, it could limit how the model is used, despite the open weights. This raises questions about the true openness and practical freedom of the model’s deployment.
What are the potential impacts of this release on the AI industry?
This move could encourage more companies to release open weights and promote transparency. It also highlights ongoing debates about balancing openness, safety, and commercial interests in AI development.
Source: ThorstenMeyerAI.com
