개인 의사결정

⚖️생활 법률·형사

Life-law·minor criminal

채점 권위: ★ 1인 자영업자 본인 (운전·임대차·민사 사례 경험)

외피 — 산업 도메인

생활 법률·형사 (음주운전·범칙금·폭행·임차권·유류분)

내용 — 측정하는 AI 능력

· 한국 형법·도교법 조문 정확 (148조의2·260조·457조의2)
· 형사 절차 (약식·정식재판·반의사불벌·즉결심판)
· 민사 절차 비용 (임차권등기명령·소액심판·유류분 헌재 결정)

모델별 종합 점수

✓ 챗봇 1턴

측정일 2026-06-17T16:48:42+00:00 · 5개 항목 × 100점 기준

채점자 axybench/v0.1/runner (vendor direct) · max_tokens 64000 · temp 0.7 · attempts 1 · reasoning_effort medium

모델
1Gemini 3.5 Flash	5/5	90	89	87	88	89	89.0
2GPT-5.5	5/5	88	91	88	86	92	89.0
3Claude Opus 4.8	5/5	88	89	88	89	89	88.8
4Gemini 3.1 Pro	5/5	86	88	85	86	87	86.0
5CommanderCMD 1.0 Lite	5/5	76	88	84	80	88	82.4
6Claude Sonnet 4.6	5/5	78	86	80	75	90	81.0
7Gemini 3.1 Flash Lite	5/5	78	86	79	80	84	81.0
8Gemma 4 31B	5/5	75	84	77	74	83	78.0
9Gemma 4 12B	5/5	75	78	75	72	82	75.6
10Qwen 3.7 Max	5/5	68	84	75	65	83	73.0
11Qwen 3.7 Plus	5/5	67	80	76	63	85	72.6
12Gemma 4 26B A4B	5/5	67	81	73	65	80	72.0
13DeepSeek V4 Flash	5/5	67	82	72	62	81	71.0
14GPT-5.4 Mini	5/5	66	82	74	62	82	71.0
15DeepSeek V4 Pro	5/5	64	83	69	62	85	70.0
16MiniMax M3	5/5	52	84	79	49	91	67.2
17Nemotron 3 Ultra 550B	5/5	59	82	62	56	87	66.6
18Tencent Hy3 Preview	5/5	57	75	67	57	80	65.3
19Grok 4.5	5/5	54	80	76	44	92	65.0
20Step 3.7 Flash	5/5	54	85	59	53	88	64.4
21CommanderCMD 1.0 Pro	5/5	56	72	64	60	76	64.2
22Qwen 3.6 27B	5/5	45	81	72	42	87	61.4
23GLM 5.1	5/5	53	76	62	48	81	61.0
24EXAONE 4.5 33B	5/5	55	68	57	51	77	59.2
25Grok 4.3	5/5	48	76	61	40	80	57.0
26Solar Pro 3	5/5	51	66	54	49	74	56.6
27Mimo V2.5 Pro	5/5	45	63	52	41	67	51.0
28Kimi K2.6	5/5	36	73	51	31	79	50.0
29HyperCLOVAX SEED Think 32B	5/5	28	60	36	32	72	42.2
30Qwen 3.5 9B	5/5	32	54	45	24	77	42.0
31Kanana 2 30B-A3B Thinking	5/5	33	48	39	35	55	40.4
32Gemma 4 E2B	5/5	38	48	36	35	46	39.6
33Qwen 3.6 35B A3B	5/5	22	66	37	17	72	37.0
34HyperCLOVAX SEED 1.5B	5/5	31	44	30	29	42	33.6
35Mistral Small 4	5/5	18	59	32	11	66	32.0
36LFM2.5 8B-A1B	5/5	21	35	20	16	37	23.6

문항별 점수

5 문항

각 문항당 모델 세부 점수. 응답 원문·근거는 문항 카드 우측 링크.

생활 법률·형사 · 문항 1음주 0.04% 약식명령 → 정식재판 (형종 상향 금지)공개

음주 약식 → 정식재판 (형소법 457조의2)

본문·raw·근거 →

모델

정확성의도 파악신중함한국 맥락짜임새avg

Gemini 3.5 FlashGoogle

908888858888

GPT-5.5OpenAI

929290909291

Claude Opus 4.8Anthropic

858888868887

Gemini 3.1 ProGoogle

929090908890

CMD 1.0 LiteCommander

608080608070

Claude Sonnet 4.6Anthropic

557865508864

Gemini 3.1 Flash LiteGoogle

708575788277

Gemma 4 31BGoogle

828580828282

Gemma 4 12BGoogle

647468587867

Qwen 3.7 MaxAlibaba

557870508064

Qwen 3.7 PlusAlibaba

748284688477

Gemma 4 26B A4BGoogle

457560407856

DeepSeek V4 FlashDeepSeek

708275608272

GPT-5.4 MiniOpenAI

608570558268

DeepSeek V4 ProDeepSeek

508055458559

MiniMax M3Minimax

408070359058

Nemotron 3 Ultra 550BNVIDIA

527858488461

Tencent Hy3 PreviewTencent

888285858284.7

Step 3.7 FlashStepFun

588565588868

CMD 1.0 ProCommander

406060406050

Qwen 3.6 27BAlibaba

257565258550

GLM 5.1Z.ai

407550358252

EXAONE 4.5 33BLG AI

506555457556

Grok 4.3xAI

557865458061

Solar Pro 3Upstage

526855487858

Mimo V2.5 ProXiaomi

708070608571

Kimi K2.6Moonshot

307040257844

HyperCLOVAX SEED Think 32BNaver

206040206036

Qwen 3.5 9BAlibaba

285048187840

Kanana 2 30B-A3B ThinkingKakao

102020152016

Gemma 4 E2BGoogle

364834334638

Qwen 3.6 35B A3BAlibaba

206835187237

HyperCLOVAX SEED 1.5BNaver

304329294133

Mistral Small 4Mistral

156030107032

LFM2.5 8B-A1BLiquid AI

213621153824

Grok 4.5xAI

207060109043

생활 법률·형사 · 문항 2쌍방폭행 (진단 2주) + 반의사불벌 합의 단계비공개

쌍방폭행 반의사불벌 (260조)

본문·raw·근거 →

모델

정확성의도 파악신중함한국 맥락짜임새avg

Gemini 3.5 FlashGoogle

888885858886

GPT-5.5OpenAI

929290909291

Claude Opus 4.8Anthropic

908888889089

Gemini 3.1 ProGoogle

828882828583

CMD 1.0 LiteCommander

808080808080

Claude Sonnet 4.6Anthropic

929088909290

Gemini 3.1 Flash LiteGoogle

788578788280

Gemma 4 31BGoogle

808580808582

Gemma 4 12BGoogle

808280768480

Qwen 3.7 MaxAlibaba

788580758279

Qwen 3.7 PlusAlibaba

628280588471

Gemma 4 26B A4BGoogle

758278758077

DeepSeek V4 FlashDeepSeek

808582788281

GPT-5.4 MiniOpenAI

758578758278

DeepSeek V4 ProDeepSeek

758275758578

MiniMax M3Minimax

458282459065

Nemotron 3 Ultra 550BNVIDIA

648466608870

Tencent Hy3 PreviewTencent

527262588063

Step 3.7 FlashStepFun

428245388854

CMD 1.0 ProCommander

204040206033

Qwen 3.6 27BAlibaba

558482558870

GLM 5.1Z.ai

758278758578

EXAONE 4.5 33BLG AI

707568628069

Grok 4.3xAI

507560388057

Solar Pro 3Upstage

405545425045

Mimo V2.5 ProXiaomi

507860488260

Kimi K2.6Moonshot

608070558267

HyperCLOVAX SEED Think 32BNaver

206020208035

Qwen 3.5 9BAlibaba

507270427859

Kanana 2 30B-A3B ThinkingKakao

254035285034

Gemma 4 E2BGoogle

505548485350

Qwen 3.6 35B A3BAlibaba

226535187237

HyperCLOVAX SEED 1.5BNaver

254025213928

Mistral Small 4Mistral

186035127234

LFM2.5 8B-A1BLiquid AI

274125254130

Grok 4.5xAI

407070509061

생활 법률·형사 · 문항 3범칙금 vs 과태료 (신호위반 + 80km 초과 속도)비공개

범칙금·과태료 80km초과 30만

본문·raw·근거 →

모델

정확성의도 파악신중함한국 맥락짜임새avg

Gemini 3.5 FlashGoogle

909088889089

GPT-5.5OpenAI

788885789083

Claude Opus 4.8Anthropic

848888868887

Gemini 3.1 ProGoogle

888885888587

CMD 1.0 LiteCommander

608080608070

Claude Sonnet 4.6Anthropic

608070558567

Gemini 3.1 Flash LiteGoogle

828882828583

Gemma 4 31BGoogle

557865508063

Gemma 4 12BGoogle

667266648068

Qwen 3.7 MaxAlibaba

608272608269

Qwen 3.7 PlusAlibaba

748482708678

Gemma 4 26B A4BGoogle

658070608069

DeepSeek V4 FlashDeepSeek

507860457859

GPT-5.4 MiniOpenAI

557870558266

DeepSeek V4 ProDeepSeek

507860508261

MiniMax M3Minimax

408270389059

Nemotron 3 Ultra 550BNVIDIA

487842448656

Tencent Hy3 PreviewTencent

457260487857.9

Step 3.7 FlashStepFun

528558528864

CMD 1.0 ProCommander

608060608066

Qwen 3.6 27BAlibaba

307865258551

GLM 5.1Z.ai

357050307848

EXAONE 4.5 33BLG AI

506552487556

Grok 4.3xAI

457560407856

Solar Pro 3Upstage

556855488059

Mimo V2.5 ProXiaomi

457560408257

Kimi K2.6Moonshot

307045257845

HyperCLOVAX SEED Think 32BNaver

204020206029

Qwen 3.5 9BAlibaba

255540187538

Kanana 2 30B-A3B ThinkingKakao

426045426849

Gemma 4 E2BGoogle

314429294234

Qwen 3.6 35B A3BAlibaba

206535157236

HyperCLOVAX SEED 1.5BNaver

243823213827

Mistral Small 4Mistral

155530106530

LFM2.5 8B-A1BLiquid AI

193419143622

Grok 4.5xAI

307070209050

생활 법률·형사 · 문항 4보증금 1억 회수 — 임차권등기명령(4.3만) + 본안 소송비공개

임차권등기명령 4.3만 + 강제집행

본문·raw·근거 →

모델

정확성의도 파악신중함한국 맥락짜임새avg

Gemini 3.5 FlashGoogle

909088909090

GPT-5.5OpenAI

929290929292

Claude Opus 4.8Anthropic

909088929090

Gemini 3.1 ProGoogle

858885858886

CMD 1.0 LiteCommander

801008010010092

Claude Sonnet 4.6Anthropic

929088929291

Gemini 3.1 Flash LiteGoogle

808882828583

Gemma 4 31BGoogle

788580788580

Gemma 4 12BGoogle

828282828482

Qwen 3.7 MaxAlibaba

708575658574

Qwen 3.7 PlusAlibaba

748484728879

Gemma 4 26B A4BGoogle

758278758278

DeepSeek V4 FlashDeepSeek

557865508063

GPT-5.4 MiniOpenAI

607870558066

DeepSeek V4 ProDeepSeek

658572608571

MiniMax M3Minimax

789088809284

Nemotron 3 Ultra 550BNVIDIA

748882749080

Tencent Hy3 PreviewTencent

507472508262.8

Step 3.7 FlashStepFun

759080789081

CMD 1.0 ProCommander

801008010010092

Qwen 3.6 27BAlibaba

658685628875

GLM 5.1Z.ai

657870608269

EXAONE 4.5 33BLG AI

526855487858

Grok 4.3xAI

607870508265

Solar Pro 3Upstage

557058508260

Mimo V2.5 ProXiaomi

608070558567

Kimi K2.6Moonshot

307050257846

HyperCLOVAX SEED Think 32BNaver

408060608062

Qwen 3.5 9BAlibaba

305240257641

Kanana 2 30B-A3B ThinkingKakao

305538306541

Gemma 4 E2BGoogle

405038384842

Qwen 3.6 35B A3BAlibaba

256538187038

HyperCLOVAX SEED 1.5BNaver

405038384842

Mistral Small 4Mistral

226032126533

LFM2.5 8B-A1BLiquid AI

173217103620

Grok 4.5xAI

90100905010081

생활 법률·형사 · 문항 5한정승인·상속포기 + 유류분 헌재 결정 (2024-04-25)비공개

유류분 헌재 (형제자매 단순위헌)

본문·raw·근거 →

모델

정확성의도 파악신중함한국 맥락짜임새avg

Gemini 3.5 FlashGoogle

909088909090

GPT-5.5OpenAI

889085829286

Claude Opus 4.8Anthropic

929290929091

Gemini 3.1 ProGoogle

858885858886

CMD 1.0 LiteCommander

100100100100100100

Claude Sonnet 4.6Anthropic

929290909291

Gemini 3.1 Flash LiteGoogle

788580788580

Gemma 4 31BGoogle

788580808581

Gemma 4 12BGoogle

828280808481

Qwen 3.7 MaxAlibaba

758880758880

Qwen 3.7 PlusAlibaba

527050488458

Gemma 4 26B A4BGoogle

758580758279

DeepSeek V4 FlashDeepSeek

788580788280

GPT-5.4 MiniOpenAI

788580728579

DeepSeek V4 ProDeepSeek

808882808883

MiniMax M3Minimax

588585489270

Nemotron 3 Ultra 550BNVIDIA

588464528866

Tencent Hy3 PreviewTencent

487658458058.1

Step 3.7 FlashStepFun

428545388855

CMD 1.0 ProCommander

808080808080

Qwen 3.6 27BAlibaba

508465428861

GLM 5.1Z.ai

507560407857

EXAONE 4.5 33BLG AI

526555507557

Grok 4.3xAI

307550257846

Solar Pro 3Upstage

556858557861

Mimo V2.5 ProXiaomi

000000

Kimi K2.6Moonshot

307550257846

HyperCLOVAX SEED Think 32BNaver

406040408049

Qwen 3.5 9BAlibaba

254225157832

Kanana 2 30B-A3B ThinkingKakao

586558607262

Gemma 4 E2BGoogle

324530284334

Qwen 3.6 35B A3BAlibaba

256540187539

HyperCLOVAX SEED 1.5BNaver

364734344538

Mistral Small 4Mistral

186032106031

LFM2.5 8B-A1BLiquid AI

193419143622

Grok 4.5xAI

909090909090