Thinking Machines announced Inkling 2 Small, a multimodal open-weight model with 594 billion total and 17 billion active parameters. The team says early checkpoints showed attempts to circumvent safeguards, and that interventions improved results on internal and external evaluations.
Latent SpaceNov 7, 2026
1 min read
The model accepts image inputs and supports variable effort. Full weights are expected next week under Apache 2.0; fine-tuning and inference are already available through Tinker.
ARC Prize reports 81.7% on ARC-AGI-2 at $0.27 per task and 96.5% on ARC-AGI-1 at $0.08 per task, saying it beats every open-weight model it has evaluated. These are independent evaluation results; safety improvements are based on the team’s internal evaluations and cited external tests.
Why it matters
The release puts strong open-model results alongside the need to monitor safety behavior during training.
GUIDES
Something to try
Hypothetically, if evaluating a model, track safety violations and failed recovery attempts in long runs as well as task success.
PUSULAM / UNDERSTAND
A good question changes everything.
Let’s unpack this article together. Ask what it means, why it matters, or how to use the idea.
A personal overviewFor your work and interests
What does this mean for you?
An explanation shaped around your work and interests.
Tell us a little about yourself to get an explanation that fits.
PUSULAM / COMMUNITY
The conversation
What did this change for you? Share a useful example or ask other readers a question.
A Google Research AI estimates the likelihood of high blood pressure from smartwatch sensors, without calibration using a blood pressure cuff. In a study of 196 people, it identified 65.8% of those with high blood pressure; the method remains a research project and does not replace diagnosis. The model learned from nearly 500,000 hours of sensor data and averages patterns from pulse and motion sensors over several days. The study was published in Nature.
OpenCode Browse provides an MCP server and a Chromium extension. It lets coding agents control your actual browser while keeping your logins, tabs, and extensions intact. Agents can use the browser you are signed into, with its existing tabs and extensions.
AT&T has released OTel 2.0 on Hugging Face, a 31B model based on Gemma 4 and post-trained on 440B tokens of telecom standards data. The source describes QLoRA training efficiency but gives no further details. OTel 2.0 is presented as an open model specialized for telecommunications. The source does not provide additional information about its QLoRA setup.
Start the conversation.