Released: 2026-09-14
Missing Domino:
meta-llama/llama-4-maverick
meta
The excluded domino is number 0 and 0
google/gemini-3-flash-preview
By cataloging all the dominoes in the image and comparing them against the full set of 28 tiles in a standard double-six set, I identified that the [6, 4] tile is the only one missing.
anthropic/claude-haiku-4.5
anthropic
By systematically cataloging all 27 visible dominoes in the image, I can identify which domino from the standard 28-tile double-six set is missing. The visible dominoes include all pairs from 0-6, including the doubles [0,0], [1,1], [2,2], [3,3], [5,5], and [6,6]. The only double missing from the complete set is [4,4], which represents four pips on both sides of a domino tile.
openai/gpt-5-image-mini
openai
After checking the pictured tiles, every domino combination from 0–6 appears except the domino containing 2 and 6; there is no tile showing a 2 on one half and a 6 on the other.
moonshotai/kimi-k2.5
moonshot
I need to identify which domino is missing from the standard 28-tile double-six set. Let me count and identify all present dominoes. The standard set has tiles: [0,0], [0,1], [0,2], [0,3], [0,4], [0,5], [0,6], [1,1], [1,2], [1,3], [1,4], [1,5], [1,6], [2,2], [2,3], [2,4], [2,5], [2,6], [3,3], [3,4], [3,5], [3,6], [4,4], [4,5], [4,6], [5,5], [5,6], [6,6]. Going through the image systematically: Row 1: [3,5] vertical, [0,2], [0,4]. Row 2: [0,5], [1,5], [1,3], [1,4], [0,3]. Row 3: [1,4], [0,0], [1,2], [3,5], [3,5] - wait let me be more careful. Looking at patterns: I can see [0,0], [0,2], [0,3], [0,4], [0,5] - missing [0,1], [0,6]. For 1s: [1,2], [1,3], [1,4], [1,5] visible, need [1,1], [1,6], [1,0]. For 2s: [2,2], [2,3], [2,4], [2,5], [2,6] checking. For 3s: [3,3], [3,4], [3,5], [3,6] etc. After careful analysis counting all visible dominoes and comparing against the complete set, the domino [1,1] (double-one) is not present in the image.
Each model receives the same domino puzzle grid. Models are tasked with identifying the missing domino using only visual clues from the grid. Their answers are evaluated without feedback or retries.