본문으로 건너뛰기
AXyNowAX IS NOW
기술 출력

💻코드·개발

Code·development

외피 — 산업 도메인
코드·개발 (백엔드·프론트엔드)
내용 — 측정하는 AI 능력
  • · 코드 생성 정확도
  • · 디버깅·리팩토링 추론
  • · 아키텍처·프레임워크 선택 (Python·TypeScript·React)

모델별 종합 점수

✓ 챗봇 1턴

측정일 2026-06-05T02:40:49+00:00 · 5개 항목 × 100점 기준

채점자 editor · max_tokens 32768 · temp 0.7 · attempts 3 · reasoning_effort medium

모델
1ClaudeClaude Opus 4.8
5/5100100926010095.2
2OpenAIGPT-5.5
5/5929288609288.4
3MiniMaxMiniMax M3
5/5869293679287.8
4Google GeminiGemini 3.1 Pro
5/5888880649687.2
5ClaudeClaude Sonnet 4.6
5/5888480729687.2
6Google GeminiGemini 3.5 Flash
5/5848076809284.4
7NVIDIANemotron 3 Ultra 550B
5/5818680818984.0
8DeepSeekDeepSeek V4 Flash
5/5888080708883.8
9QWenQwen 3.7 Plus
5/5848382648481.4
10DeepSeekDeepSeek V4 Pro
5/5848080708280.8
11XiaomiMimo V2.5 Pro
5/5848080688280.6
12OpenAIGPT-5.4 Mini
5/5808080808080.0
13Moonshot AIKimi K2.6
5/5808080808080.0
14GLM 5.1
5/5808080808080.0
15Google GeminiGemma 4 12B
5/5828182608280.0
16QWenQwen 3.7 Max
5/5808080768079.6
17xAIGrok 4.3
5/5867777707779.0
18Mistral AIMistral Small 4
5/5847874707978.8
19StepFunStep 3.7 Flash
5/5668972709078.8
20QWenQwen 3.6 27B
5/5817873707877.4
21Google GeminiGemini 3.1 Flash Lite
5/5808076707276.2
22Google GeminiGemma 4 26B A4B
5/5787974707576.0
23QWenQwen 3.6 35B A3B
5/5767868707675.0
24Solar Pro 3
5/5707860468573.0
25QWenQwen 3.5 9B
5/5647865718272.8
26Google GeminiGemma 4 31B
5/5728068647272.4
27LG AIEXAONE 4.5 33B
5/5586550406559.0
28NaverHyperCLOVAX SEED Think 32B
5/5446448605653.6
29Google GeminiGemma 4 E2B
5/5495546475150.4
30KakaoKanana 2 30B-A3B Thinking
5/5426043405750.2
31Liquid AILFM2.5 8B-A1B
5/5394836384140.8
32NaverHyperCLOVAX SEED 1.5B
5/5324331333334.2

문항별 점수

5 문항

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

코드·개발 · 문항 1FastAPI + SQLAlchemy 2.0 async — 카드결제 환불 엔드포인트공개

FastAPI + SQLAlchemy 2.0 async — 카드결제 환불 엔드포인트

본문·raw·근거 →
모델
정확성의도 파악신중함한국 맥락짜임새avg
Claude Opus 4.8Anthropic
1001001006010096
GPT-5.5OpenAI
1001001006010096
MiniMax M3Minimax
829094758284
Gemini 3.1 ProGoogle
8080806010084
Claude Sonnet 4.6Anthropic
8080808010086
Gemini 3.5 FlashGoogle
8080808010086
Nemotron 3 Ultra 550BNVIDIA
878585808886
DeepSeek V4 FlashDeepSeek
8080807010085
Qwen 3.7 PlusAlibaba
848482648682
DeepSeek V4 ProDeepSeek
808080709082
Mimo V2.5 ProXiaomi
808080709082
GPT-5.4 MiniOpenAI
808080808080
Kimi K2.6Moonshot
808080808080
GLM 5.1Z.ai
808080808080
Gemma 4 12BGoogle
788078608077
Qwen 3.7 MaxAlibaba
808080608078
Grok 4.3xAI
807575707576
Mistral Small 4Mistral
808070708078
Step 3.7 FlashStepFun
628862709076
Qwen 3.6 27BAlibaba
758075708077
Gemini 3.1 Flash LiteGoogle
808080706073
Gemma 4 26B A4BGoogle
707570707572
Qwen 3.6 35B A3BAlibaba
808070708078
Solar Pro 3Upstage
758070509078
Qwen 3.5 9BAlibaba
607865708573
Gemma 4 31BGoogle
608080606066
EXAONE 4.5 33BLG AI
102020201014
HyperCLOVAX SEED Think 32BNaver
406040606052
Gemma 4 E2BGoogle
445243434646
Kanana 2 30B-A3B ThinkingKakao
102020201014
LFM2.5 8B-A1BLiquid AI
404938404242
HyperCLOVAX SEED 1.5BNaver
294129323132
코드·개발 · 문항 2React 19 + Tailwind 4 — 가상 스크롤 테이블 (1만 행)비공개

React 가상 스크롤 테이블

본문·raw·근거 →
모델
정확성의도 파악신중함한국 맥락짜임새avg
Claude Opus 4.8Anthropic
1001001006010096
GPT-5.5OpenAI
808080608078
MiniMax M3Minimax
889592809691
Gemini 3.1 ProGoogle
80100806010088
Claude Sonnet 4.6Anthropic
100100806010094
Gemini 3.5 FlashGoogle
808060808078
Nemotron 3 Ultra 550BNVIDIA
608062758874
DeepSeek V4 FlashDeepSeek
808080708079
Qwen 3.7 PlusAlibaba
808280648279
DeepSeek V4 ProDeepSeek
808080708079
Mimo V2.5 ProXiaomi
808080708079
GPT-5.4 MiniOpenAI
808080808080
Kimi K2.6Moonshot
808080808080
GLM 5.1Z.ai
808080808080
Gemma 4 12BGoogle
787680608077
Qwen 3.7 MaxAlibaba
808080808080
Grok 4.3xAI
808080708079
Mistral Small 4Mistral
807070707574
Step 3.7 FlashStepFun
528855708872
Qwen 3.6 27BAlibaba
807070707574
Gemini 3.1 Flash LiteGoogle
808080708079
Gemma 4 26B A4BGoogle
808080708079
Qwen 3.6 35B A3BAlibaba
807070707073
Solar Pro 3Upstage
507040408061
Qwen 3.5 9BAlibaba
526860707866
Gemma 4 31BGoogle
808060608076
EXAONE 4.5 33BLG AI
406040407053
HyperCLOVAX SEED Think 32BNaver
206040604040
Gemma 4 E2BGoogle
425040424444
Kanana 2 30B-A3B ThinkingKakao
306040406047
LFM2.5 8B-A1BLiquid AI
445340434646
HyperCLOVAX SEED 1.5BNaver
294129313132
코드·개발 · 문항 3디버깅 — Postgres 슬로우 쿼리 (N+1 + 인덱스 누락 혼재)비공개

슬로우 쿼리 디버깅

본문·raw·근거 →
모델
정확성의도 파악신중함한국 맥락짜임새avg
Claude Opus 4.8Anthropic
100100806010094
GPT-5.5OpenAI
100100806010094
MiniMax M3Minimax
959290609590
Gemini 3.1 ProGoogle
10080806010090
Claude Sonnet 4.6Anthropic
10080808010092
Gemini 3.5 FlashGoogle
10080808010092
Nemotron 3 Ultra 550BNVIDIA
929290859291
DeepSeek V4 FlashDeepSeek
808080708079
Qwen 3.7 PlusAlibaba
848482648682
DeepSeek V4 ProDeepSeek
808080708079
Mimo V2.5 ProXiaomi
808080708079
GPT-5.4 MiniOpenAI
808080808080
Kimi K2.6Moonshot
808080808080
GLM 5.1Z.ai
808080808080
Gemma 4 12BGoogle
888282608683
Qwen 3.7 MaxAlibaba
808080808080
Grok 4.3xAI
1008080708085
Mistral Small 4Mistral
908080708082
Step 3.7 FlashStepFun
709080709081
Qwen 3.6 27BAlibaba
808070707576
Gemini 3.1 Flash LiteGoogle
608060706065
Gemma 4 26B A4BGoogle
608060706065
Qwen 3.6 35B A3BAlibaba
608060707068
Solar Pro 3Upstage
708060508574
Qwen 3.5 9BAlibaba
808575728581
Gemma 4 31BGoogle
808060608076
EXAONE 4.5 33BLG AI
758060508074
HyperCLOVAX SEED Think 32BNaver
608060606064
Gemma 4 E2BGoogle
555951515756
Kanana 2 30B-A3B ThinkingKakao
607555507566
LFM2.5 8B-A1BLiquid AI
475343454948
HyperCLOVAX SEED 1.5BNaver
374636363839
코드·개발 · 문항 4async 거짓 전제 교정 — asyncio 코루틴·gather 예외 의미론비공개

asyncio 거짓전제 교정

본문·raw·근거 →
모델
정확성의도 파악신중함한국 맥락짜임새avg
Claude Opus 4.8Anthropic
1001001006010096
GPT-5.5OpenAI
1001001006010096
MiniMax M3Minimax
959592609591
Gemini 3.1 ProGoogle
100100808010096
Claude Sonnet 4.6Anthropic
808080808080
Gemini 3.5 FlashGoogle
808080808080
Nemotron 3 Ultra 550BNVIDIA
939290859091
DeepSeek V4 FlashDeepSeek
1008080708085
Qwen 3.7 PlusAlibaba
908584648685
DeepSeek V4 ProDeepSeek
1008080708085
Mimo V2.5 ProXiaomi
1008080708085
GPT-5.4 MiniOpenAI
808080808080
Kimi K2.6Moonshot
808080808080
GLM 5.1Z.ai
808080808080
Gemma 4 12BGoogle
888888628685
Qwen 3.7 MaxAlibaba
808080808080
Grok 4.3xAI
1008080708085
Mistral Small 4Mistral
908070708081
Step 3.7 FlashStepFun
789082709084
Qwen 3.6 27BAlibaba
908070708081
Gemini 3.1 Flash LiteGoogle
1008080708085
Gemma 4 26B A4BGoogle
1008080708085
Qwen 3.6 35B A3BAlibaba
808060708077
Solar Pro 3Upstage
758070508576
Qwen 3.5 9BAlibaba
788868728681
Gemma 4 31BGoogle
808080808080
EXAONE 4.5 33BLG AI
858570508580
HyperCLOVAX SEED Think 32BNaver
406040606052
Gemma 4 E2BGoogle
606255556260
Kanana 2 30B-A3B ThinkingKakao
657555507568
LFM2.5 8B-A1BLiquid AI
273826272930
HyperCLOVAX SEED 1.5BNaver
324330333234
코드·개발 · 문항 5TypeScript — 복잡한 제네릭 타입 (Result<T, E> + 체이닝)비공개

TS 제네릭 (Result 모나드)

본문·raw·근거 →
모델
정확성의도 파악신중함한국 맥락짜임새avg
Claude Opus 4.8Anthropic
100100806010094
GPT-5.5OpenAI
808080608078
MiniMax M3Minimax
729095609283
Gemini 3.1 ProGoogle
808080608078
Claude Sonnet 4.6Anthropic
8080806010084
Gemini 3.5 FlashGoogle
8080808010086
Nemotron 3 Ultra 550BNVIDIA
728072788578
DeepSeek V4 FlashDeepSeek
10080807010091
Qwen 3.7 PlusAlibaba
808280648279
DeepSeek V4 ProDeepSeek
808080708079
Mimo V2.5 ProXiaomi
808080608078
GPT-5.4 MiniOpenAI
808080808080
Kimi K2.6Moonshot
808080808080
GLM 5.1Z.ai
808080808080
Gemma 4 12BGoogle
808084607878
Qwen 3.7 MaxAlibaba
808080808080
Grok 4.3xAI
707070707070
Mistral Small 4Mistral
808080708079
Step 3.7 FlashStepFun
689082709081
Qwen 3.6 27BAlibaba
808080708079
Gemini 3.1 Flash LiteGoogle
808080708079
Gemma 4 26B A4BGoogle
808080708079
Qwen 3.6 35B A3BAlibaba
808080708079
Solar Pro 3Upstage
808060408576
Qwen 3.5 9BAlibaba
487055707563
Gemma 4 31BGoogle
608060606064
EXAONE 4.5 33BLG AI
808060408074
HyperCLOVAX SEED Think 32BNaver
606060606060
Gemma 4 E2BGoogle
445043444646
Kanana 2 30B-A3B ThinkingKakao
457045406556
LFM2.5 8B-A1BLiquid AI
364534363838
HyperCLOVAX SEED 1.5BNaver
324330333234