SciCode

16개 과학 분야의 80개 실험실 문제에서 가져온 288개 테스트 세트 하위 문제로 구성된, 과학자가 큐레이션한 코딩 벤치마크입니다.

모델: 142 현재, 251 전체
중앙값: 48.7%
범위: 6.9%–67.7%

점수 분포

가장 낮음: 6.9%
중앙값: 48.7%
최고: 67.7%

리더보드

지원 중단된 모델 포함

순위	모델	점수
1	Claude Fable 5.2 (Max, Default Fallback) Anthropic	67.7%
2	Claude Opus 5.5 (Max, Default Fallback) Anthropic	66.9%
3	Claude Fable 5.2 (Xhigh, Default Fallback) Anthropic	65.6%
4	Claude Opus 5.5 (Xhigh, Default Fallback) Anthropic	65.0%
5	Muse Wonder Full (Max) Meta	62.6%
6	Gemini 4 Argon (High) Google	61.8%
7	Claude Sonnet 5.5 (Max, Default Fallback) Anthropic	61.0%
8	MiMo-V2.6-Pro Xiaomi	60.9%
9	Kimi K3.1 Pro Kimi	60.8%
10	Claude Opus 5.5 (High, Default Fallback) Anthropic	60.4%
11	Muse Wonder (Max) Meta	60.3%
12	Claude Fable 5.2 (Medium, Default Fallback) Anthropic	59.7%
13	Claude Fable 5.2 (High, Default Fallback) Anthropic	59.5%
14	Claude Opus 5.5 (Medium, Default Fallback) Anthropic	59.3%
15	GLM-5.3 (Max) Z AI	59.0%
16	Step 5 Preview StepFun	58.9%
17	Gemini 3.1 Pro Preview Google	58.7%
18	Claude Opus 5.5 (Low, Default Fallback) Anthropic	58.6%
19	Muse Spark Open (Xhigh) Meta	58.4%
20	Claude Fable 5.2 (Low, Default Fallback) Anthropic	57.8%
21	Grok 4.7 (High) SpaceXAI	57.8%
22	Grok 4.7 (Xhigh) SpaceXAI	57.4%
23	Claude Sonnet 5.5 (Xhigh, Default Fallback) Anthropic	57.3%
24	Ling 3.1 Flash 1019 InclusionAI	56.8%
25	Gemini 3.8 Flash (High) Google	56.6%
26	GPT-6 Astra (Max) OpenAI	56.5%
27	Agnes 2.5 Pro Theta Sapiens AI	56.0%
28	DeepSeek V4.1 Flash 1017 (Max) DeepSeek	55.9%
29	GPT-6.1 Sol (High) OpenAI	55.8%
30	GPT-6.1 Sol (Xhigh) OpenAI	55.7%
31	GPT-6 Astra (Xhigh) OpenAI	55.7%
32	GPT-6 Astra (High) OpenAI	55.4%
33	Gemini 3.8 Flash (Medium) Google	55.1%
34	GPT-6.1 Terra (Max) OpenAI	55.1%
35	Claude Haiku 5.5 (Max) Anthropic	55.0%
36	Gemini 3.8 Flash (Low) Google	55.0%
37	Grok 4.7 (Low) SpaceXAI	54.9%
38	GPT-6 Luna (Max) OpenAI	54.6%
39	GPT-6.1 Sol (Max) OpenAI	54.2%
40	GPT-6 Astra (Medium) OpenAI	54.2%
41	GPT-6 Astra (Low) OpenAI	54.1%
42	Ling 3.1 Flash InclusionAI	54.1%
43	Qwen3.8 2.4T A95B Alibaba	54.1%
44	Claude Sonnet 5.5 (High, Default Fallback) Anthropic	53.7%
45	GPT-6.1 Terra (Xhigh) OpenAI	53.4%
46	MiniMax-M3.1-Flash MiniMax	53.4%
47	GPT-6.1 Sol (Low) OpenAI	53.2%
48	GPT-6.1 Sol (Medium) OpenAI	53.2%
49	Claude Sonnet 5.5 (Medium, Default Fallback) Anthropic	52.9%
50	GPT-5.5 Instant (June 2026) OpenAI	52.5%

1
Claude Fable 5.2 (Max, Default Fallback)
Anthropic
67.7%
2
Claude Opus 5.5 (Max, Default Fallback)

데이터 출처: Artificial Analysis

Command Palette

SciCode

점수 분포

리더보드