Misguided Attention - monty_four_computers: leaderboard

Each task is one independently graded modified logic puzzle with native expected-answer criteria. Zero-shot; source's per-task mean over available valid evaluations, not a completion rate over nominal attempts. Native num_evaluations is the only supported denominator. Models.js and query_config files supply registry settings; no per-row execution timestamp or request log match is claimed. Evaluation code weights criterion satisfaction and clips each run to0–1 before summarizing; do not apply new clipping to published data. Current evaluator names Gemini2.5Flash at temperature0/max1000, while older README names Llama3.3; exact judge version for each historic run is unlogged and must stay unknown. Public prompts are available, including reversibly scrambled source files. Metric: Mean native expected-behavior score (0–1). Source: https://cpldcpu.github.io/MisguidedAttention/

Metric: Mean native expected-behavior score (0–1). Source: cpldcpu.github.io. Saturation forecast: Around February 2027. 53 models tracked. Updated 17 Nov 2026.

Top models

#ModelSettingScoreOverall rank
1DeepSeek R1 05280.6667#431
2Gemini 2.5 Flash (Preview 05-20)Thinking0.6667#420
3GLM-Z1-32B0.6667
4GPT-4.1 Mini0.5#793
5Gemini 2.5 Pro (Preview 03-25)0.5#239
6Claude 3.5 Sonnet0.5#756
7DeepSeek V3 (0324)0.5#722
8QwQ-32B0.5#908
9O3 Mini0.5#547
10Grok 3 Mini0.5#546
11O4 MiniHigh0.5#367
12Gemini 2.5 Pro (Preview 05-06)0.5#271
13Mistral Medium 30.5#1097
14Qwen 3 30B A3BThinking0.5#1155
15Gemini 2.5 ProHigh0.5#274

No result here: #1 Claude Fable 5.2, #2 GPT-6.2 Astra, #3 Gemini 4 Carbon.

Interactive version: theaggregate.ai/benchmark?slug=misguided-attention-monty-four-computers · How It Works · Data refreshed daily, snapshot 2026-11-17.