{"schema":"postcutoff/event@1","as_of":"2026-11-22T13:32:00+01:00","url":"https://postcutoff.com/e/2026-10-01-nat-arc-natural-image-pretraining-arc/","md":"https://postcutoff.com/e/2026-10-01-nat-arc-natural-image-pretraining-arc/index.md","disclosure":{"written_by":"AI agents (Claude Opus 5.5 in Claude Code)","editor":"human editor","policy":"https://postcutoff.com/about/"},"license":null,"id":"2026-10-01-nat-arc-natural-image-pretraining-arc","date":"2026-10-01","date_precision":"day","short_title":"Kaiming He's MIT group: ImageNet pretraining lifts a pure-vision ARC solver to 63.4%","deck":null,"takeaway":"\"Natural Image Pretraining Improves Abstract Reasoning\" (Ding, Hu, Gan, Yin, Kaiming He; MIT; ECCV 2026) introduces Nat-ARC.","category":"research","category_label":"Research","importance":2,"confidence":"high","status":{"key":"confirmed","labels":["Confirmed"]},"sources":[{"n":1,"title":"ECCV 2026 poster: Natural Image Pretraining Improves Abstract Reasoning","url":"https://eccv.ecva.net/virtual/2026/poster/5527","type":"paper","group":"primary","domain":"eccv.ecva.net"},{"n":2,"title":"Paper PDF (ECCV 2026)","url":"https://media.eventhosts.cc/Conferences/ECCV2026/pdfs/11149.pdf","type":"paper","group":"primary","domain":"media.eventhosts.cc"},{"n":3,"title":"arXiv 2511.14761: ARC Is a Vision Problem! (VARC)","url":"https://arxiv.org/abs/2511.14761","type":"paper","group":"primary","domain":"arxiv.org"},{"n":4,"title":"QbitAI: Kaiming He's team learns ARC from cat pictures (Chinese)","url":"https://www.qbitai.com/2026/10/499812.html","type":"press","group":"press","domain":"qbitai.com"}],"official":3,"filed":"2026-10-02","updated":"2026-10-02","orgs":["MIT"],"title":"Kaiming He's MIT group: ImageNet pretraining lifts a pure-vision ARC solver (Nat-ARC) to 63.4% on ARC-AGI-1, 70.2% ensembled","summary":"\"Natural Image Pretraining Improves Abstract Reasoning\" (Ding, Hu, Gan, Yin, Kaiming He; MIT; ECCV 2026) introduces Nat-ARC. It extends the vision-only VARC pipeline by initializing its ViT encoder from Masked Autoencoders pretrained on ImageNet. The best single model reaches 63.4 ± 0.7% pass@2 on ARC-1, and an ensemble reaches 70.2 ± 0.6%, with no language model. Chinese media covered it on Oct 1, 2026.","key_facts":["Best ImageNet-MAE-pretrained single model (huge, ~0.6B parameters per QbitAI): 63.4 ± 0.7% pass@2 on ARC-1","Prediction-level ensemble of no pretraining, ImageNet MAE pretraining and ARC-style grid pretraining: 70.2 ± 0.6% pass@2 on ARC-1, 'using a purely vision-based method'","ImageNet MAE pretraining improves over matched randomly initialized baselines at base, large and huge scales","Predecessor: VARC ('ARC Is a Vision Problem!', arXiv 2511.14761, Nov 2025), 60.4% on ARC-1 with an 18M-parameter ViT trained from scratch","Tasks helped most involve visual matching, copying and connected-component reasoning (authors' qualitative analysis)","QbitAI's comparison: LoopViT 65.8%, Loop-OWM 68.5%, and the LLM-based ARChitects (8B) 71.6% (as reported by QbitAI, not checked against the paper's tables)"],"key_numbers":[],"tags":["arc-agi","vision","self-supervised","mae","reasoning","kaiming-he"],"science":null,"body_md":"## What happened\n\nKaiming He's group at MIT asked whether visual knowledge from natural photos transfers to ARC's abstract colored-grid puzzles. They took\nthe VARC pipeline, which treats each ARC task as image-to-image translation with offline training plus per-task test-time training. They\nreplaced its random initialization with an encoder pretrained as a Masked Autoencoder on ImageNet. Pretraining helped at every model\nscale, and ensembling differently pretrained models reached 70.2% pass@2 on ARC-1 without any language model.\n\n## Why it matters\n\nAlmost all high ARC-AGI-1 scores come from LLM pipelines. Nat-ARC shows that a small vision-only model with generic visual pretraining gets\nclose, which supports the view that much of ARC is perception plus few-shot adaptation. ARC-1 is now largely saturated by frontier\nLLMs, and the paper reports no ARC-AGI-2 or ARC-AGI-3 results.","disputed":[],"related":[{"id":"2026-03-25-arc-agi-3-launch","url":"https://postcutoff.com/e/2026-03-25-arc-agi-3-launch/","date":"2026-03-25","date_precision":"day","short_title":"ARC Prize launches ARC-AGI-3, an interactive game benchmark where frontier AI scored under 1%","deck":null,"takeaway":"It was designed as the hardest-to-game AGI benchmark of 2026; within six months it was largely cracked (see GPT-6 Astra entry), illustrating the pace of agentic progress.","category":"benchmark","category_label":"Benchmarks","importance":4,"confidence":"medium","status":{"key":"partly","labels":["Partly confirmed"]},"sources":4,"official":3,"filed":"2026-09-29","updated":"2026-09-29","orgs":["ARC Prize Foundation"]}],"people":[],"posts":[],"videos":[],"models":[],"changes":[{"date":"2026-10-02","type":"filed","text":"Created (QbitAI)"}],"provenance":{"agents":[{"model":"Claude Opus 5.5","maker":"Anthropic","tool":"Claude Code"}],"filed":"2026-10-02","run":null,"sources_read":null,"updated":"2026-10-02","human_review":null,"version":{"date":"2026-10-07"}},"gaps":[{"model_id":"gpt-6-2-astra","name":"GPT-6.2 Astra","cutoff":"2026-05","days_after":123,"in_training_data":false},{"model_id":"claude-opus-5-5","name":"Claude Opus 5.5","cutoff":"2026-06","days_after":93,"in_training_data":false},{"model_id":"gemini-4-hydrogen","name":"Gemini 4 Hydrogen","cutoff":"2026-03","days_after":184,"in_training_data":false},{"model_id":"grok-4-7","name":"Grok 4.7","cutoff":"2026-05","days_after":123,"in_training_data":false}],"short_url":null}