LMGame-Bench Ace Attorney Leaderboard — The Aggregate

LMGame-Bench Ace Attorney — leaderboard

LLM game-playing benchmark: Ace Attorney testing logical deduction and evidence-based reasoning in courtroom scenarios.

Metric: Score. Source: huggingface.co. Status: saturation imminent. 17 models tracked.

Top models

#	Model	Score
1	O3 (2025-04-16)	16
2	O1 (2024-12-17)	16
3	GPT-5 (High)	9
4	Claude 3.7 Sonnet (20250219)	7
5	Gemini 2.5 Pro (Preview 05-06)	7
6	Claude Opus 4 (20250514)	6
7	O4 Mini (2025-04-16)	4
8	Gemini 2.5 Flash (Preview 04-17)	4
9	Claude Sonnet 4 (20250514)	3.7
10	Claude 3.5 Sonnet (20241022)	2
11	GPT-4.1	2
12	DeepSeek R1	0
13	GPT-4o (2024-11-20)	0
14	Llama 4 Maverick Instruct FP8	0
15	O1 Mini (2024-09-12)	0

Interactive version: theaggregate.ai/benchmark?slug=lmgame-bench-ace-attorney · How the rankings work · Data refreshed daily, snapshot 2026-07-22.