Mira Murati's Inkling Tops Every Open-Weight Model on ARC-AGI-2
Thinking Machines' Inkling becomes the top open-weight model on ARC-AGI-1 and ARC-AGI-2, scoring 79.5% and 36.5% respectively at under $1 per task.
- Inkling by Thinking Machines Lab is now the highest-scoring open-weight model on both ARC-AGI-1 (79.5%) and ARC-AGI-2 (36.5%), verified by ARC Prize.
- Costs are $0.30/task on ARC-AGI-1 and $0.64/task on ARC-AGI-2, making it competitive on efficiency as well as accuracy.
- Inkling is a 975B-parameter Mixture-of-Experts model (41B active) with a 1M token context window, pretrained on 45 trillion tokens of text, images, audio, and video.
- Full weights are available on Hugging Face under Apache 2.0; fine-tuning is available via Thinking Machines' Tinker platform at a 50% introductory discount.
- A smaller variant, Inkling-Small (276B total, 12B active), matches or beats the full model on several benchmarks; full weights coming soon.
- ARC-AGI-3 evaluation for Inkling is pending; closed-weight frontier models (GPT-5.5 at 85%) still lead ARC-AGI-2 by a wide margin.
Thinking Machines Lab, the AI startup founded by former OpenAI CTO Mira Murati, just got its first major benchmark crown. ARC Prize has officially verified that Inkling, Thinking Machines' debut open-weight model, is now the highest-scoring open-weight model on both ARC-AGI-1 and ARC-AGI-2. It hits 79.5% on ARC-AGI-1 and 36.5% on ARC-AGI-2, at costs of $0.30 and $0.64 per task respectively. For context, that ARC-AGI-2 score puts it ahead of every other open-weight model that has been formally evaluated by ARC Prize, in a benchmark where frontier closed models like GPT-5.5 sit around 85%.
What ARC-AGI actually tests
ARC-AGI is not a typical benchmark. Each puzzle ships with 2-8 examples, and the model must infer a hidden rule (color swaps, symmetry, recursion, compositional transforms) and apply it to a withheld test grid, with no memorized facts, just pattern discovery and reasoning. It is deliberately designed to resist the kind of pattern-matching that large training corpora enable. Average individual human performance on ARC-AGI-2 is 66%. The benchmark is graded with zero tolerance: answers must match ground-truth output exactly, any discrepancy in length, structure, or formatting fails the task, and no partial credit is given, with attempts limited to three per test input.
ARC-AGI-2 launched in early 2025 with every frontier model at 0%. Progress over the past year has been steep. The open-weight side of the leaderboard, however, has lagged far behind closed-weight systems, making Inkling's 36.5% a meaningful data point for the open-source community.
What Inkling actually is
Thinking Machines Lab released its first in-house model, and unlike the flagships from OpenAI, Anthropic, or Google, anyone can download it. Inkling is a Mixture-of-Experts transformer with 975 billion total parameters and 41 billion active, a 1 million token context window, and 45 trillion tokens of pretraining across text, images, audio and video. The MoE design (Mixture-of-Experts) means the model routes each token through a small subset of its parameters, keeping inference fast and cheap despite the massive total parameter count.