Misguided Attention - monty_defective_appliance: 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: Not forecast. 53 models tracked. Updated 17 Nov 2026.

Top models

#ModelSettingScoreOverall rank
1Llama 3.3 70B Instruct0.6667#1387
2Gemini 2.5 Flash (Preview 05-20)0.6667#420
3Gemini 2.5 FlashHigh0.6667#467
4GPT-4o0.5#754
5Claude 3.7 Sonnet0.5#476
6Grok 30.5#606
7Grok 3 Mini Beta0.5#620
8O3 Mini0.5#547
9O4 MiniHigh0.5#367
10Mistral Medium 30.5#1097
11Optimus Alpha0.5#526
12GLM-Z1-32B0.5
13GPT-4.1 Mini0.3333#793
14GPT-4.10.3333#505
15Gemini 2.0 Flash0.3333#723

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-defective-appliance · How It Works · Data refreshed daily, snapshot 2026-11-17.