순위표

추론 랭킹

심층 추론과 과학: GPQA Diamond, HLE, ARC-AGI-2.

종합700 코딩399 수학53 추론413 에이전트55 지식412 멀티모달382 코딩 에이전트55 과학 추론82 다국어12 글쓰기61 채팅 어시스턴트61 리서치85 도구 호출32 긴 컨텍스트82 속도402 가성비401 오픈 웨이트170

413개 모델

#	모델명	Podium Score	GPQA Diamond	Humanity’s Last Exam	ARC-AGI-2	속도
01	Claude Mythos Preview Anthropic · reasoning · preview	98.8	94.6%	64.7%	—	80 t/s
02	GPT-6 Astra OpenAI · reasoning	96.0	96%	53.2%	—	82 t/s
03	Claude Fable 5.1 Anthropic · reasoning	92.0	—	—	—	85 t/s
4	GPT-5.6 Sol OpenAI	87.6	94.6%	47.2%	—	72 t/s
5	Kimi K3 Moonshot AI · 오픈소스 · reasoning	86.5	93.5%	56%	—	37 t/s
6	Claude Opus 4.8 Anthropic	82.8	93.6%	57.9%	—	50 t/s
7	Claude Opus 5 Anthropic	82.3	93.2%	52.6%	—	60 t/s
8	Claude Fable 5 Anthropic · reasoning	81.0	92.6%	53.3%	—	71 t/s
9	GPT-5.5 OpenAI	80.4	93.6%	52.2%	85%	60 t/s
10	Gemini 3.1 Pro Google	73.5	94.3%	51.4%	77.1%	136 t/s
11	Claude Opus 4.7 Anthropic	73.3	94.2%	54.7%	—	80 t/s
12	Claude Mythos 5 Anthropic · reasoning · preview	71.0	—	—	—	50 t/s
13	GPT-5.6 Terra OpenAI	69.6	92.9%	41.8%	—	127 t/s
14	Gemini 3.5 Flash Google	66.5	92.2%	41%	—	267 t/s
15	Qwen3.8 Max Alibaba	66.3	92.6%	43.6%	—	55 t/s
16	Qwen3.8 Flash Next Alibaba · 오픈소스	66.0	81%	—	—	168 t/s
17	Claude Opus 4.6 Anthropic	66.0	91.3%	53.1%	68.8%	50 t/s
18	Gemini 3.6 Flash Google	65.7	92.8%	38.3%	—	233 t/s
19	GLM-5.2 Z.ai · 오픈소스	65.6	91.2%	54.7%	—	164 t/s
20	MiniMax M3 MiniMax	64.8	92.9%	37.1%	—	93 t/s
21	GPT-5.3 Codex OpenAI	62.9	91.5%	39.9%	—	118 t/s
22	Muse Spark 1.1 Meta	62.7	89.8%	45.1%	—	208 t/s
23	GLM-5.3 Flash Z.ai · 오픈소스	62.5	78.5%	—	—	182 t/s
24	GLM-5.1 Z.ai · 오픈소스	62.0	67%	—	—	74 t/s
25	Claude Sonnet 4.6 (max) Anthropic	62.0	67%	—	—	50 t/s
26	Qwen3.7 Max Alibaba	61.0	92.4%	41.4%	—	203 t/s
27	Grok 4.5 xAI	60.8	93%	40.3%	—	59 t/s
28	GPT-5.4 OpenAI	60.1	92.8%	39.8%	73.3%	60 t/s
29	MiniMax-M2.7 MiniMax	60.0	64%	—	—	40 t/s
30	DeepSeek V4.1 Flash DeepSeek · 오픈소스 · reasoning	58.0	62%	—	—	214 t/s
31	Gemini 3.8 Flash Google · reasoning	56.0	—	—	—	302 t/s
32	Gemini 3.8 Flash (Medium Reasoning) Google · reasoning	56.0	—	—	—	315 t/s
33	Gemini 3.8 Flash (Low Reasoning) Google · reasoning	56.0	—	—	—	340 t/s
34	GPT-5.1 OpenAI	55.5	88.1%	—	—	60 t/s
35	GPT-5.6 Luna OpenAI	55.1	92.3%	37.2%	—	175 t/s
36	Claude Opus 4 7 Non Reasoning Anthropic · reasoning	55.0	—	—	—	50 t/s
37	Claude Sonnet 5 Non Reasoning Anthropic · reasoning	55.0	—	—	—	50 t/s
38	Claude Opus 4 5 Thinking Anthropic · reasoning	55.0	—	—	—	50 t/s
39	Gpt 5 6 Sol Non Reasoning OpenAI · reasoning	55.0	—	—	—	60 t/s
40	Solar Pro4 Upstage · reasoning	55.0	—	—	—	40 t/s
41	Deepseek V3 2 Reasoning DeepSeek · 오픈소스 · reasoning	55.0	—	—	—	40 t/s
42	Claude 4 5 Sonnet Thinking Anthropic · reasoning	55.0	—	—	—	50 t/s
43	Glm 5 1 Non Reasoning Z.ai · 오픈소스 · reasoning	55.0	—	—	—	40 t/s
44	Gpt 5 5 Non Reasoning OpenAI · reasoning	55.0	—	—	—	60 t/s
45	Kimi K2 6 Non Reasoning Moonshot AI · reasoning	55.0	—	—	—	40 t/s
46	Claude Sonnet 4 6 Non Reasoning Low Effort Anthropic · reasoning	55.0	—	—	—	50 t/s
47	Glm 5 2 Non Reasoning Z.ai · 오픈소스 · reasoning	55.0	—	—	—	40 t/s
48	Gpt 5 6 Terra Non Reasoning OpenAI · reasoning	55.0	—	—	—	60 t/s
49	Claude 4 1 Opus Thinking Anthropic · reasoning	55.0	—	—	—	50 t/s
50	Kimi K2 Thinking Moonshot AI · reasoning	55.0	—	—	—	40 t/s
51	O3 Pro OpenAI · reasoning	55.0	—	—	—	60 t/s
52	Glm 5 Non Reasoning Z.ai · 오픈소스 · reasoning	55.0	—	—	—	40 t/s
53	Claude 4 5 Haiku Reasoning Anthropic · reasoning	55.0	—	—	—	50 t/s
54	Qwen3 5 397b A17b Non Reasoning Alibaba · 오픈소스 · reasoning	55.0	—	—	—	55 t/s
55	Nova 2 0 Pro Reasoning Medium Amazon · reasoning	55.0	—	—	—	40 t/s
56	Deepseek V4 Pro 0424 Non Reasoning DeepSeek · 오픈소스 · reasoning	55.0	—	—	—	40 t/s
57	Claude 4 Opus Thinking Anthropic · reasoning	55.0	—	—	—	50 t/s
58	Qwen3 6 27b Non Reasoning Alibaba · 오픈소스 · reasoning	55.0	—	—	—	55 t/s
59	Deepseek V3 1 Terminus Reasoning DeepSeek · 오픈소스 · reasoning	55.0	—	—	—	40 t/s
60	Kimi K2 5 Non Reasoning Moonshot AI · reasoning	55.0	—	—	—	40 t/s
61	Qwen3 5 27b Non Reasoning Alibaba · 오픈소스 · reasoning	55.0	—	—	—	55 t/s
62	Claude 4 Sonnet Thinking Anthropic · reasoning	55.0	—	—	—	50 t/s
63	Glm 4 6 Reasoning Z.ai · 오픈소스 · reasoning	55.0	—	—	—	40 t/s
64	Mimo V2 5 Pro Non Reasoning Xiaomi · 오픈소스 · reasoning	55.0	—	—	—	40 t/s
65	Gpt 5 4 Non Reasoning OpenAI · reasoning	55.0	—	—	—	60 t/s
66	Qwen3 5 122b A10b Non Reasoning Alibaba · 오픈소스 · reasoning	55.0	—	—	—	55 t/s
67	Grok 4 Fast Reasoning xAI · reasoning	55.0	—	—	—	50 t/s
68	Glm 4 7 Non Reasoning Z.ai · 오픈소스 · reasoning	55.0	—	—	—	40 t/s
69	Gpt 5 6 Luna Non Reasoning OpenAI · reasoning	55.0	—	—	—	60 t/s
70	Hy3 Non Reasoning Tencent · reasoning	55.0	—	—	—	40 t/s
71	Gpt 5 2 Non Reasoning OpenAI · reasoning	55.0	—	—	—	60 t/s
72	Deepseek V3 2 Reasoning 0925 DeepSeek · 오픈소스 · reasoning	55.0	—	—	—	40 t/s
73	Qwen3 Max Thinking Preview Alibaba · 오픈소스 · reasoning	55.0	—	—	—	55 t/s
74	Grok 4 3 Non Reasoning xAI · reasoning	55.0	—	—	—	50 t/s
75	Qwen3 6 35b A3b Non Reasoning Alibaba · 오픈소스 · reasoning	55.0	—	—	—	55 t/s
76	Qwen3 5 35b A3b Non Reasoning Alibaba · 오픈소스 · reasoning	55.0	—	—	—	55 t/s
77	O1 OpenAI · reasoning	55.0	—	—	—	60 t/s
78	Grok 4 20 0309 Non Reasoning xAI · reasoning	55.0	—	—	—	50 t/s
79	Gemma 4 31b Non Reasoning Google · 오픈소스 · reasoning	55.0	—	—	—	80 t/s
80	Grok 4 20 Non Reasoning xAI · reasoning	55.0	—	—	—	50 t/s
81	Nova 2 0 Omni Reasoning Medium Amazon · reasoning	55.0	—	—	—	40 t/s
82	Deepseek V3 1 Reasoning DeepSeek · 오픈소스 · reasoning	55.0	—	—	—	40 t/s
83	Qwen3 Vl 235b A22b Reasoning Alibaba · 오픈소스 · reasoning	55.0	—	—	—	55 t/s
84	Gpt 5 1 Non Reasoning OpenAI · reasoning	55.0	—	—	—	60 t/s
85	Qwen3 5 9b Non Reasoning Alibaba · 오픈소스 · reasoning	55.0	—	—	—	55 t/s
86	Gemma 4 26b A4b Non Reasoning Google · 오픈소스 · reasoning	55.0	—	—	—	80 t/s
87	Qwen3 235b A22b Instruct 2507 Reasoning Alibaba · 오픈소스 · reasoning	55.0	—	—	—	55 t/s
88	Nova 2 0 Pro Reasoning Low Amazon · reasoning	55.0	—	—	—	40 t/s
89	O1 Pro OpenAI · reasoning	55.0	—	—	—	60 t/s
90	Deepseek R1 0120 DeepSeek · 오픈소스 · reasoning	55.0	—	—	—	40 t/s
91	Trinity Large Thinking Trinity · reasoning	55.0	—	—	—	40 t/s
92	Qwen3 Vl 32b Reasoning Alibaba · 오픈소스 · reasoning	55.0	—	—	—	55 t/s
93	O1 Preview OpenAI · reasoning	55.0	—	—	—	60 t/s
94	Glm 4 6v Reasoning Z.ai · 오픈소스 · reasoning	55.0	—	—	—	40 t/s
95	Qwen3 Next 80b A3b Reasoning Alibaba · 오픈소스 · reasoning	55.0	—	—	—	55 t/s
96	Nova 2 0 Omni Reasoning Low Amazon · reasoning	55.0	—	—	—	40 t/s
97	Qwen3 5 4b Non Reasoning Alibaba · 오픈소스 · reasoning	55.0	—	—	—	55 t/s
98	Qwen3 30b A3b 2507 Reasoning Alibaba · 오픈소스 · reasoning	55.0	—	—	—	55 t/s
99	Qwen3 235b A22b Instruct Reasoning Alibaba · 오픈소스 · reasoning	55.0	—	—	—	55 t/s
100	Qwen3 Vl 30b A3b Reasoning Alibaba · 오픈소스 · reasoning	55.0	—	—	—	55 t/s
101	Gemma 4 12b Non Reasoning Google · 오픈소스 · reasoning	55.0	—	—	—	80 t/s
102	Llama Nemotron Super 49b V1 5 Reasoning Meta · 오픈소스 · reasoning	55.0	—	—	—	50 t/s
103	Mistral Small 4 Non Reasoning Mistral · 오픈소스 · reasoning	55.0	—	—	—	60 t/s
104	Qwen3 32b Instruct Reasoning Alibaba · 오픈소스 · reasoning	55.0	—	—	—	55 t/s
105	Hermes 4 Llama 3 1 70b Reasoning Meta · 오픈소스 · reasoning	55.0	—	—	—	50 t/s
106	Deepseek R1 Distill Llama 70b DeepSeek · 오픈소스 · reasoning	55.0	—	—	—	40 t/s
107	Qwen3 Omni 30b A3b Reasoning Alibaba · 오픈소스 · reasoning	55.0	—	—	—	55 t/s
108	Qwen3 30b A3b Instruct Reasoning Alibaba · 오픈소스 · reasoning	55.0	—	—	—	55 t/s
109	Glm 4 5v Reasoning Z.ai · 오픈소스 · reasoning	55.0	—	—	—	40 t/s
110	Llama 3 1 Nemotron Ultra 253b V1 Reasoning Meta · 오픈소스 · reasoning	55.0	—	—	—	50 t/s
111	Hermes 4 Llama 3 1 405b Reasoning Meta · 오픈소스 · reasoning	55.0	—	—	—	50 t/s
112	Gemma 4 E4b Non Reasoning Google · 오픈소스 · reasoning	55.0	—	—	—	80 t/s
113	Cogito V2 1 Reasoning Cogito · reasoning	55.0	—	—	—	40 t/s
114	Hidream O1 Image OpenAI · reasoning	55.0	—	—	—	60 t/s
115	Granite 4.2 30B Instruct IBM · 오픈소스	54.5	68%	—	—	78 t/s
116	DeepSeek V4 Pro DeepSeek · 오픈소스	54.5	90.1%	48.2%	—	66 t/s
117	Kimi K2.7 Code Moonshot AI · 오픈소스	48.9	89.6%	32.8%	—	39 t/s
118	Kimi K2.6 Moonshot AI · 오픈소스	48.7	90.5%	36.4%	—	40 t/s
119	Agnes 3.0 Flash Agnes · reasoning	48.0	49%	—	—	235 t/s
120	Claude Opus 4.5 Anthropic	47.9	87%	—	—	50 t/s
121	Muse Spark Meta	47.3	89.5%	58.4%	42.5%	80 t/s
122	Mistral Medium 3.5 Mistral	47.0	51%	—	—	60 t/s
123	Claude Sonnet 4.6 Anthropic	46.7	89.9%	49%	58.3%	50 t/s
124	Granite 4.2 8B Instruct IBM · 오픈소스	46.0	54%	—	—	115 t/s
125	Qwen3.7 Plus Alibaba	45.6	90.3%	34.7%	—	54 t/s
126	Qwen3.5 122B-A10B Alibaba · 오픈소스	45.2	86.6%	—	—	126 t/s
127	DeepSeek V4 Flash Vision DeepSeek	45.0	36%	—	—	120 t/s
128	G9v3-39A5B AI9Stars · 오픈소스	45.0	36%	—	—	60 t/s
129	G9v3-3B AI9Stars · 오픈소스	45.0	36%	—	—	60 t/s
130	GLM-5.3 Zhipu AI	45.0	36%	—	—	90 t/s
131	Grok 4.6 (low) xAI	45.0	36%	—	—	59 t/s
132	Grok 4.6 (medium) xAI	45.0	36%	—	—	59 t/s
133	Grok 4.6 (xhigh) xAI	45.0	36%	—	—	62 t/s
134	HyperNova 60B Multiverse · 오픈소스	45.0	36%	—	—	367 t/s
135	LFM2.5-2.6B Liquid AI · 오픈소스	45.0	36%	—	—	60 t/s
136	LFM2.5-8B-A1B Liquid AI · 오픈소스	45.0	36%	—	—	338 t/s
137	LongCat 2.0 LongCat · 오픈소스	45.0	36%	—	—	42 t/s
138	Magistral Small 1.2 Mistral · 오픈소스	45.0	36%	—	—	60 t/s
139	Qwen3.8 27B (xhigh) Alibaba · 오픈소스	45.0	36%	—	—	53 t/s
140	Qwen3.8 27B (low) Alibaba · 오픈소스	45.0	36%	—	—	64 t/s
141	Qwen3.8 27B (medium) Alibaba · 오픈소스	45.0	36%	—	—	62 t/s
142	Qwen3.8 27B (Non-reasoning) Alibaba · 오픈소스 · reasoning	45.0	36%	—	—	62 t/s
143	Claude 4.5 Haiku Anthropic	45.0	48%	—	—	80 t/s
144	GPT-5.5 Pro OpenAI	45.0	—	—	—	60 t/s
145	Grok 4 Heavy xAI · preview	44.9	88.4%	50.7%	—	50 t/s
146	GPT-5.2 OpenAI	44.3	92.4%	34.5%	52.9%	90 t/s
147	Nova 2.0 Pro Preview Amazon · preview	43.0	46%	—	—	40 t/s
148	Hunyuan Hy3 Tencent	42.7	90.4%	31.6%	—	69 t/s
149	Agnes 2.5 Pro Beta Agnes	41.5	45%	—	—	65 t/s
150	DeepSeek V4 Flash DeepSeek · 오픈소스	41.2	88.1%	45.1%	—	113 t/s
151	GPT-oss-120B OpenAI · 오픈소스	40.0	43%	—	—	240 t/s
152	MiMo V2.5 Pro Xiaomi · 오픈소스	39.7	86.6%	33.8%	—	65 t/s
153	GPT-5 OpenAI	39.0	85.7%	—	—	60 t/s
154	Granite 4.2 3B Instruct IBM · 오픈소스	38.0	42%	—	—	148 t/s
155	Inkling Thinking Machines	37.3	87.2%	29.7%	—	81 t/s
156	Qwen3.6 Plus Alibaba	36.9	90.4%	28.8%	—	55 t/s
157	Gemini 3 Pro Google	36.2	91.9%	45.8%	31.1%	120 t/s
158	Claude Opus 5 Xhigh Anthropic	35.0	—	—	—	80 t/s
159	Claude Opus 5 High Anthropic	35.0	—	—	—	80 t/s
160	Grok 4 xAI	35.0	—	—	—	80 t/s
161	Grok 4 6 xAI	35.0	—	—	—	80 t/s
162	Claude Opus 5 Medium Anthropic	35.0	—	—	—	80 t/s
163	Qwen3 8 2 4t A95b Alibaba · 오픈소스	35.0	—	—	—	85 t/s
164	Gpt 5 6 Sol High OpenAI	35.0	—	—	—	90 t/s
165	Muse Spark 1 2 Meta	35.0	—	—	—	80 t/s
166	Gpt 5 6 Sol Medium OpenAI	35.0	—	—	—	90 t/s
167	Glm 5 Z.ai · 오픈소스	35.0	—	—	—	60 t/s
168	Claude Opus 5 Low Anthropic	35.0	—	—	—	80 t/s
169	Gpt 5 5 Medium OpenAI	35.0	—	—	—	90 t/s
170	Gpt 5 6 Sol Low OpenAI	35.0	—	—	—	90 t/s
171	Gpt 5 6 Terra High OpenAI	35.0	—	—	—	90 t/s
172	Kimi K3 Low Moonshot AI · reasoning	35.0	—	—	—	60 t/s
173	Claude Sonnet 4 6 Adaptive Anthropic	35.0	—	—	—	80 t/s
174	Gpt 5 6 Luna High OpenAI	35.0	—	—	—	90 t/s
175	Gpt 5 6 Terra Medium OpenAI	35.0	—	—	—	90 t/s
176	Deepseek V4 Pro 0424 DeepSeek · 오픈소스	35.0	—	—	—	60 t/s
177	Claude Opus 4 6 Adaptive Anthropic	35.0	—	—	—	80 t/s
178	Gpt 5 5 Low OpenAI	35.0	—	—	—	90 t/s
179	Deepseek V4 Pro 0424 High DeepSeek · 오픈소스	35.0	—	—	—	60 t/s
180	Gpt 5 6 Terra Low OpenAI	35.0	—	—	—	90 t/s
181	Gpt 5 2 Codex OpenAI	35.0	—	—	—	90 t/s
182	Qwen3 6 Max Alibaba · 오픈소스	35.0	—	—	—	85 t/s
183	Grok Build 0 1 06 16 xAI	35.0	—	—	—	80 t/s
184	Gpt 5 4 Low OpenAI	35.0	—	—	—	90 t/s
185	Agnes 2 5 Pro Alpha Agnes	35.0	—	—	—	60 t/s
186	Gpt 5 2 Medium OpenAI	35.0	—	—	—	90 t/s
187	Gpt 5 6 Luna Medium OpenAI	35.0	—	—	—	90 t/s
188	Gemini 3 Flash Reasoning Google · reasoning	35.0	—	—	—	80 t/s
189	Nvidia Nemotron 3 Ultra 550b A55b NVIDIA	35.0	—	—	—	150 t/s
190	Grok 4 20 xAI	35.0	—	—	—	80 t/s
191	Mimo V2 5 0424 Xiaomi · 오픈소스	35.0	—	—	—	60 t/s
192	Qwen3 6 27b Alibaba · 오픈소스	35.0	—	—	—	85 t/s
193	Grok 4 20 0309 xAI	35.0	—	—	—	80 t/s
194	Gpt 5 Codex OpenAI	35.0	—	—	—	90 t/s
195	Grok 4 3 Medium xAI	35.0	—	—	—	80 t/s
196	Grok 4 3 Low xAI	35.0	—	—	—	80 t/s
197	Kimi K2 5 Moonshot AI	35.0	—	—	—	60 t/s
198	Gpt 5 1 Codex OpenAI	35.0	—	—	—	90 t/s
199	Muse Glimmer Meta	35.0	—	—	—	80 t/s
200	Gpt 5 Medium OpenAI	35.0	—	—	—	90 t/s
201	Hy3 Preview Tencent	35.0	—	—	—	60 t/s
202	Gpt 5 5 Instant 05 26 OpenAI	35.0	—	—	—	90 t/s
203	Gpt 5 6 Luna Low OpenAI	35.0	—	—	—	90 t/s
204	Qwen3 6 35b A3b Alibaba · 오픈소스	35.0	—	—	—	85 t/s
205	Nvidia Nemotron 3 Super 120b A12b NVIDIA	35.0	—	—	—	150 t/s
206	Gpt 5 Low OpenAI	35.0	—	—	—	90 t/s
207	Mimo V2 Flash Reasoning Xiaomi · 오픈소스 · reasoning	35.0	—	—	—	40 t/s
208	Qwen3 5 Omni Plus Alibaba · 오픈소스	35.0	—	—	—	85 t/s
209	Ring 2 6 1t WeRide	35.0	—	—	—	60 t/s
210	Claude 4 5 Sonnet Anthropic	35.0	—	—	—	80 t/s
211	Deepseek V4 Flash Non Reasoning DeepSeek · 오픈소스 · reasoning	35.0	—	—	—	40 t/s
212	Gpt 5 5 Instant 06 26 OpenAI	35.0	—	—	—	90 t/s
213	Claude 4 1 Opus Anthropic	35.0	—	—	—	80 t/s
214	Ling 2 6 1t Ling	35.0	—	—	—	60 t/s
215	Gemma 4 26b A4b Google · 오픈소스	35.0	—	—	—	120 t/s
216	Claude 4 Opus Anthropic	35.0	—	—	—	80 t/s
217	Claude 4 Sonnet Anthropic	35.0	—	—	—	80 t/s
218	Ling 3 0 Tiny Ling	35.0	—	—	—	60 t/s
219	Kimi K2 0905 Moonshot AI	35.0	—	—	—	60 t/s
220	Claude 3 7 Sonnet Anthropic	35.0	—	—	—	80 t/s
221	Nemotron 3 5 Lightning NVIDIA	35.0	—	—	—	150 t/s
222	Grok 3 Mini Reasoning xAI · reasoning	35.0	—	—	—	50 t/s
223	Command A Plus Cohere	35.0	—	—	—	60 t/s
224	Gemma 4 12b Google · 오픈소스	35.0	—	—	—	120 t/s
225	Mercury 2 Inception	35.0	—	—	—	60 t/s
226	Qwen3 5 9b Alibaba · 오픈소스	35.0	—	—	—	85 t/s
227	Deepseek V3 2 0925 DeepSeek · 오픈소스	35.0	—	—	—	60 t/s
228	Apriel V1 5 15b Thinker ServiceNow	35.0	—	—	—	60 t/s
229	Qwen3 Coder Next Alibaba · 오픈소스	35.0	—	—	—	85 t/s
230	Apriel V1 6 15b Thinker ServiceNow	35.0	—	—	—	60 t/s
231	Nova 2 0 Lite Reasoning Amazon · reasoning	35.0	—	—	—	40 t/s
232	Qwen3 5 4b Alibaba · 오픈소스	35.0	—	—	—	85 t/s
233	Gemini 2 5 Flash Reasoning Google · reasoning	35.0	—	—	—	80 t/s
234	Kimi K2 Moonshot AI	35.0	—	—	—	60 t/s
235	Mistral Small 4 Mistral · 오픈소스	35.0	—	—	—	100 t/s
236	Gpt 4 1 OpenAI	35.0	—	—	—	90 t/s
237	Qwen3 Max Preview Alibaba · 오픈소스	35.0	—	—	—	85 t/s
238	Devstral 2 Mistral	35.0	—	—	—	100 t/s
239	Nova 2 0 Lite Reasoning Medium Amazon · reasoning	35.0	—	—	—	40 t/s
240	O3 Mini OpenAI · reasoning	35.0	—	—	—	60 t/s
241	Seed Oss 36b Instruct ByteDance	35.0	—	—	—	60 t/s
242	Qwen3 235b A22b Instruct 2507 Alibaba · 오픈소스	35.0	—	—	—	85 t/s
243	Qwen3 Coder 480b A35b Instruct Alibaba · 오픈소스	35.0	—	—	—	85 t/s
244	Nova 2 0 Lite Reasoning Low Amazon · reasoning	35.0	—	—	—	40 t/s
245	Gpt 5 4 Nano Non Reasoning OpenAI · reasoning	35.0	—	—	—	60 t/s
246	Devstral Small 2 Mistral	35.0	—	—	—	100 t/s
247	Gpt 5 4 Mini Non Reasoning OpenAI · reasoning	35.0	—	—	—	60 t/s
248	Grok 4 Fast xAI	35.0	—	—	—	80 t/s
249	Mistral Large 3 Mistral · 오픈소스	35.0	—	—	—	100 t/s
250	O3 Mini High OpenAI · reasoning	35.0	—	—	—	60 t/s
251	Glm 4 7 Flash Non Reasoning Z.ai · 오픈소스 · reasoning	35.0	—	—	—	40 t/s
252	Deepseek V3 0324 DeepSeek · 오픈소스	35.0	—	—	—	60 t/s
253	Gemini 2 5 Flash Lite Preview 09 2025 Reasoning Google · reasoning	35.0	—	—	—	80 t/s
254	Gpt Oss 120b Low OpenAI	35.0	—	—	—	90 t/s
255	Mistral Small 3 1 Mistral · 오픈소스	35.0	—	—	—	100 t/s
256	Mistral Medium 3 1 Mistral · 오픈소스	35.0	—	—	—	100 t/s
257	Nvidia Nemotron 3 Nano 30b A3b Reasoning NVIDIA · reasoning	35.0	—	—	—	100 t/s
258	Gpt Oss 20b Low OpenAI	35.0	—	—	—	90 t/s
259	Qwen3 Vl 235b A22b Instruct Alibaba · 오픈소스	35.0	—	—	—	85 t/s
260	Qwen3 Next 80b A3b Instruct Alibaba · 오픈소스	35.0	—	—	—	85 t/s
261	Qwen3 Coder 30b A3b Instruct Alibaba · 오픈소스	35.0	—	—	—	85 t/s
262	Qwq 32b Alibaba · reasoning	35.0	—	—	—	85 t/s
263	Nova Premier Amazon	35.0	—	—	—	60 t/s
264	Mistral Medium 3 Mistral · 오픈소스	35.0	—	—	—	100 t/s
265	Celeris 1 Celeris	35.0	—	—	—	60 t/s
266	Gemma 4 E4b Google · 오픈소스	35.0	—	—	—	120 t/s
267	Claude 3 Opus Anthropic	35.0	—	—	—	80 t/s
268	Gemini 2 5 Flash Lite Reasoning Google · reasoning	35.0	—	—	—	80 t/s
269	Qwen3 Vl 32b Instruct Alibaba · 오픈소스	35.0	—	—	—	85 t/s
270	Qwen3 235b A22b Instruct Alibaba · 오픈소스	35.0	—	—	—	85 t/s
271	Mistral Small 3 2 Mistral · 오픈소스	35.0	—	—	—	100 t/s
272	Qwen3 Vl 8b Reasoning Alibaba · 오픈소스 · reasoning	35.0	—	—	—	55 t/s
273	Nova 2 0 Omni Amazon	35.0	—	—	—	60 t/s
274	Qwen3 14b Instruct Reasoning Alibaba · 오픈소스 · reasoning	35.0	—	—	—	55 t/s
275	Llama 4 Scout Meta · 오픈소스	35.0	—	—	—	80 t/s
276	Qwen3 Vl 30b A3b Instruct Alibaba · 오픈소스	35.0	—	—	—	85 t/s
277	Claude 35 Sonnet Anthropic	35.0	—	—	—	80 t/s
278	Gpt 4o 2024 08 06 OpenAI	35.0	—	—	—	90 t/s
279	Llama 3 3 Instruct 70b Meta · 오픈소스	35.0	—	—	—	80 t/s
280	Ernie 4 5 300b A47b Baidu	35.0	—	—	—	60 t/s
281	Qwen3 30b A3b 2507 Alibaba · 오픈소스	35.0	—	—	—	85 t/s
282	Nvidia Nemotron Nano 12b V2 Vl Reasoning NVIDIA · reasoning	35.0	—	—	—	100 t/s
283	Nvidia Nemotron Nano 9b V2 Reasoning NVIDIA · reasoning	35.0	—	—	—	100 t/s
284	Hermes 4 Llama 3 1 405b Meta · 오픈소스	35.0	—	—	—	80 t/s
285	Llama Nemotron Super 49b V1 5 Meta · 오픈소스	35.0	—	—	—	80 t/s
286	Qwen3 32b Instruct Alibaba · 오픈소스	35.0	—	—	—	85 t/s
287	Gpt 4o 2024 05 13 OpenAI	35.0	—	—	—	90 t/s
288	Qwen3 8b Instruct Reasoning Alibaba · 오픈소스 · reasoning	35.0	—	—	—	55 t/s
289	Claude 35 Sonnet June 24 Anthropic	35.0	—	—	—	80 t/s
290	Gpt 4 Turbo OpenAI	35.0	—	—	—	90 t/s
291	Command A Cohere	35.0	—	—	—	60 t/s
292	Nova Pro Amazon	35.0	—	—	—	60 t/s
293	Gemma 3 27b Google · 오픈소스	35.0	—	—	—	120 t/s
294	Llama 3 1 Nemotron Instruct 70b Meta · 오픈소스	35.0	—	—	—	80 t/s
295	Mistral Large 2407 Mistral · 오픈소스	35.0	—	—	—	100 t/s
296	Gpt 4 OpenAI	35.0	—	—	—	90 t/s
297	Hermes 4 Llama 3 1 70b Meta · 오픈소스	35.0	—	—	—	80 t/s
298	Mistral Small 3 Mistral · 오픈소스	35.0	—	—	—	100 t/s
299	Qwen3 30b A3b Instruct Alibaba · 오픈소스	35.0	—	—	—	85 t/s
300	Llama 3 1 Instruct 70b Meta · 오픈소스	35.0	—	—	—	80 t/s
301	Qwen Turbo Alibaba · 오픈소스	35.0	—	—	—	85 t/s
302	Gemma 3 12b Google · 오픈소스	35.0	—	—	—	120 t/s
303	Granite 4 0 H Small IBM · 오픈소스	35.0	—	—	—	60 t/s
304	Hermes 3 Llama 3 1 70b Meta · 오픈소스	35.0	—	—	—	80 t/s
305	Jamba 1 5 Large AI21 Labs	35.0	—	—	—	60 t/s
306	Qwen3 Omni 30b A3b Instruct Alibaba · 오픈소스	35.0	—	—	—	85 t/s
307	Nova Micro Amazon	35.0	—	—	—	60 t/s
308	Mistral Small Mistral · 오픈소스	35.0	—	—	—	100 t/s
309	Mistral Large Mistral · 오픈소스	35.0	—	—	—	100 t/s
310	Claude 3 Haiku Anthropic	35.0	—	—	—	80 t/s
311	Gpt 35 Turbo OpenAI	35.0	—	—	—	90 t/s
312	Mistral Medium Mistral · 오픈소스	35.0	—	—	—	100 t/s
313	Mistral Small 2402 Mistral · 오픈소스	35.0	—	—	—	100 t/s
314	Llama 3 Instruct 70b Meta · 오픈소스	35.0	—	—	—	80 t/s
315	Llama 3 2 Instruct 11b Vision Meta · 오픈소스	35.0	—	—	—	80 t/s
316	Command R Plus 04 2024 Cohere	35.0	—	—	—	60 t/s
317	Apertus 70b Instruct Swiss AI	35.0	—	—	—	60 t/s
318	Mixtral 8x7b Instruct Mistral	35.0	—	—	—	100 t/s
319	Command R 03 2024 Cohere	35.0	—	—	—	60 t/s
320	Gemma 3 4b Google · 오픈소스	35.0	—	—	—	120 t/s
321	Gemma 3n E2b Google · 오픈소스	35.0	—	—	—	120 t/s
322	Gemma 3n E4b Google · 오픈소스	35.0	—	—	—	120 t/s
323	Tiny Aya Global Cohere	35.0	—	—	—	60 t/s
324	Amazon Nova Amazon	35.0	—	—	—	60 t/s
325	Async Pro V1 0 Async	35.0	—	—	—	60 t/s
326	Asyncflow V2 Async	35.0	—	—	—	60 t/s
327	Claude Sonnet 5 High Anthropic	35.0	—	—	—	80 t/s
328	Claude Sonnet 5 Low Anthropic	35.0	—	—	—	80 t/s
329	Claude Sonnet 5 Medium Anthropic	35.0	—	—	—	80 t/s
330	Claude Sonnet 5 Xhigh Anthropic	35.0	—	—	—	80 t/s
331	Falcon 2 TII · 오픈소스	35.0	—	—	—	60 t/s
332	Gladia Solaria Upstage	35.0	—	—	—	60 t/s
333	Gpt 5 4 Pro OpenAI	35.0	—	—	—	90 t/s
334	Gpt Realtime 2 OpenAI	35.0	—	—	—	90 t/s
335	Grok Stt xAI	35.0	—	—	—	80 t/s
336	Mistral Mistral · 오픈소스	35.0	—	—	—	100 t/s
337	Nvidia NVIDIA	35.0	—	—	—	150 t/s
338	Nvidia Nemotron Asr Streaming NVIDIA	35.0	—	—	—	150 t/s
339	Ocateve 2 Ocat	35.0	—	—	—	60 t/s
340	Qwen Image Alibaba · 오픈소스	35.0	—	—	—	85 t/s
341	Qwen3 Asr Alibaba · 오픈소스	35.0	—	—	—	85 t/s
342	Seed 2.0 Pro ByteDance	34.2	88.9%	—	—	40 t/s
343	Nemotron 3 Ultra 550B NVIDIA · 오픈소스	33.4	87%	26.6%	—	142 t/s
344	Gemini 3 Flash Google	32.7	90.4%	43.5%	33.6%	120 t/s
345	Solar Pro 3 Upstage	31.0	34%	—	—	40 t/s
346	Qwen3.5 397B-A17B Alibaba · 오픈소스	29.0	88.4%	28.7%	—	71 t/s
347	GPT-oss-20B OpenAI · 오픈소스	29.0	32%	—	—	90 t/s
348	K2 Think V2 K2	29.0	31%	—	—	40 t/s
349	Gemini 3 7 Flash Google	20.0	—	—	—	120 t/s
350	Gemini 3 7 Flash Medium Google	20.0	—	—	—	120 t/s
351	Gemini 3 7 Flash Low Google	20.0	—	—	—	120 t/s
352	Gpt 5 4 Mini OpenAI	20.0	—	—	—	90 t/s
353	Deepseek V4 Flash 0420 DeepSeek · 오픈소스	20.0	—	—	—	60 t/s
354	Gpt 5 4 Nano OpenAI	20.0	—	—	—	90 t/s
355	Deepseek V4 Flash 0420 High DeepSeek · 오픈소스	20.0	—	—	—	60 t/s
356	Ling 3 0 Flash Ling	20.0	—	—	—	60 t/s
357	Gemini 3 5 Flash Minimal Google	20.0	—	—	—	120 t/s
358	Gemini 3 Pro Low Google	20.0	—	—	—	120 t/s
359	Minimax M2 1 MiniMax · 오픈소스	20.0	—	—	—	60 t/s
360	Gpt 5 Mini Medium OpenAI	20.0	—	—	—	90 t/s
361	Gpt 5 1 Codex Mini OpenAI	20.0	—	—	—	90 t/s
362	Step 3 7 Flash StepFun	20.0	—	—	—	60 t/s
363	Gpt 5 4 Nano Medium OpenAI	20.0	—	—	—	90 t/s
364	Gpt 5 4 Mini Medium OpenAI	20.0	—	—	—	90 t/s
365	Minimax M2 MiniMax · 오픈소스	20.0	—	—	—	60 t/s
366	Step 3 5 Flash 0202 StepFun	20.0	—	—	—	60 t/s
367	Gpt 5 Mini OpenAI	20.0	—	—	—	90 t/s
368	Gemini 2 5 Pro 05 06 Google	20.0	—	—	—	120 t/s
369	North Mini Code North	20.0	—	—	—	60 t/s
370	Gpt 5 Nano OpenAI	20.0	—	—	—	90 t/s
371	Gpt 5 Nano Medium OpenAI	20.0	—	—	—	90 t/s
372	Qwen3 5 Omni Flash Alibaba · 오픈소스	20.0	—	—	—	85 t/s
373	Minimax M1 80k MiniMax · 오픈소스	20.0	—	—	—	60 t/s
374	Gpt 5 Minimal OpenAI	20.0	—	—	—	90 t/s
375	Nemotron 3 Nano Omni 30b A3b NVIDIA	20.0	—	—	—	150 t/s
376	Gpt 4 1 Mini OpenAI	20.0	—	—	—	90 t/s
377	Gpt 5 Mini Minimal OpenAI	20.0	—	—	—	90 t/s
378	Ling 2 6 Flash Ling	20.0	—	—	—	60 t/s
379	Gemini 2 5 Flash Lite Preview 09 2025 Google	20.0	—	—	—	120 t/s
380	Nova 2 0 Lite Amazon	20.0	—	—	—	60 t/s
381	Ministral 3 14b Mistral	20.0	—	—	—	100 t/s
382	Gemini 2 0 Flash Experimental Google	20.0	—	—	—	120 t/s
383	Gpt 4 1 Nano OpenAI	20.0	—	—	—	90 t/s
384	Ministral 3 8b Mistral	20.0	—	—	—	100 t/s
385	Qwen3 Vl 8b Instruct Alibaba · 오픈소스	20.0	—	—	—	85 t/s
386	Gpt 5 Nano Minimal OpenAI	20.0	—	—	—	90 t/s
387	Llama 3 1 Instruct 8b Meta · 오픈소스	20.0	—	—	—	80 t/s
388	Nvidia Nemotron 3 Nano 30b A3b NVIDIA	20.0	—	—	—	150 t/s
389	Nvidia Nemotron Nano 9b V2 NVIDIA	20.0	—	—	—	150 t/s
390	Ministral 3 3b Mistral	20.0	—	—	—	100 t/s
391	Qwen3 14b Instruct Alibaba · 오픈소스	20.0	—	—	—	85 t/s
392	Gemini 2 5 Flash Lite Google	20.0	—	—	—	120 t/s
393	Nova Lite Amazon	20.0	—	—	—	60 t/s
394	Reka Flash Reka	20.0	—	—	—	60 t/s
395	Solar Mini Upstage	20.0	—	—	—	60 t/s
396	Phi 4 Mini Microsoft · 오픈소스	20.0	—	—	—	60 t/s
397	Qwen3 8b Instruct Alibaba · 오픈소스	20.0	—	—	—	85 t/s
398	Nvidia Nemotron Nano 12b V2 Vl NVIDIA	20.0	—	—	—	150 t/s
399	Llama 2 Chat 7b Meta · 오픈소스	20.0	—	—	—	80 t/s
400	Reka Flash 3 Reka	20.0	—	—	—	60 t/s
401	Olmo 3 7b Instruct AI2 · 오픈소스	20.0	—	—	—	60 t/s
402	Jamba 1 5 Mini AI21 Labs	20.0	—	—	—	60 t/s
403	Mistral 7b Instruct Mistral · 오픈소스	20.0	—	—	—	100 t/s
404	Granite 3 3 8b Instruct IBM · 오픈소스	20.0	—	—	—	60 t/s
405	Apertus 8b Instruct Swiss AI	20.0	—	—	—	60 t/s
406	Gemma 3 1b Google · 오픈소스	20.0	—	—	—	120 t/s
407	Llama 3 Instruct 8b Meta · 오픈소스	20.0	—	—	—	80 t/s
408	Async Flash V1 5 Async	20.0	—	—	—	60 t/s
409	Gemini Google	20.0	—	—	—	120 t/s
410	Gemini 2 5 Flash Dec 2025 Google	20.0	—	—	—	120 t/s
411	Gemini_nano Banana Pro Google	20.0	—	—	—	120 t/s
412	Minimax H3 MiniMax · 오픈소스	20.0	—	—	—	60 t/s
413	Claude Sonnet 5 Anthropic	0.0	80%	17.8%	—	96 t/s

Podium Score는 5개 리더보드의 가중 평균입니다. 자세한 내용은 평가 방법을 참고하세요.