LLM Leaderboard

Frontier LLMs ranked

Name: LLM Leaderboard — Frontier model rankings
Creator: Artificial Analysis
License: https://artificialanalysis.ai/

Industry-standard rankings by Intelligence, Coding, and Math indices. Sourced from Artificial Analysis. Click into any model on Opper for full benchmarks, hosting routes, and privacy posture.

220 models

Model directory

#	Model	Intelligence	Coding	Math	Speed	On Opper
1	Claude Opus 4.8 (Adaptive Reasoning, Max Effort)Anthropic	61.4	56.7	—	55 tok/s	View →
2	GPT-5.5 (xhigh)OpenAI	60.2	59.1	—	75 tok/s	View →
3	GPT-5.5 (high)OpenAI	58.9	58.5	—	60 tok/s	View →
4	Claude Opus 4.7 (Adaptive Reasoning, Max Effort)Anthropic	57.3	52.5	—	51 tok/s	View →
5	Gemini 3.1 Pro PreviewGoogle	57.2	55.5	—	123 tok/s	View →
6	GPT-5.4 (xhigh)OpenAI	56.8	57.2	—	80 tok/s	View →
7	GPT-5.5 (medium)OpenAI	56.7	56.2	—	54 tok/s	View →
8	Qwen3.7 MaxAlibaba	56.6	50.1	—	171 tok/s	View →
9	Gemini 3.5 Flash (high)Google	55.3	45.0	—	198 tok/s	View →
10	Gemini 3.5 Flash (medium)Google	54.8	43.9	—	206 tok/s	View →
11	MiniMax-M3MiniMax	54.7	43.4	—	—	—
12	Kimi K2.6Moonshot	53.9	47.1	—	42 tok/s	View →
13	MiMo-V2.5-ProXiaomi	53.8	45.5	—	44 tok/s	—
14	GPT-5.3 Codex (xhigh)OpenAI	53.6	53.1	—	76 tok/s	View →
15	Qwen3.7 PlusAlibaba	53.3	46.5	—	54 tok/s	—
16	Grok 4.3 (high)xAI	53.2	41.0	—	123 tok/s	View →
17	Claude Opus 4.6 (Adaptive Reasoning, Max Effort)Anthropic	52.9	48.1	—	49 tok/s	View →
18	Muse SparkMeta	52.2	47.5	—	—	—
19	Claude Opus 4.7 (Non-reasoning, High Effort)Anthropic	51.8	53.1	—	42 tok/s	View →
20	Qwen3.6 Max PreviewAlibaba	51.8	44.9	—	41 tok/s	—
21	Claude Sonnet 4.6 (Adaptive Reasoning, Max Effort)Anthropic	51.7	50.9	—	66 tok/s	View →
22	DeepSeek V4 Pro (Reasoning, Max Effort)DeepSeek	51.5	47.5	—	50 tok/s	View →
23	GLM-5.1 (Reasoning)Z.ai	51.4	43.4	—	50 tok/s	View →
24	GPT-5.2 (xhigh)OpenAI	51.3	48.7	99.0	71 tok/s	View →
25	GPT-5.5 (low)OpenAI	50.8	52.1	—	52 tok/s	View →
26	Qwen3.6 PlusAlibaba	50.0	42.9	—	53 tok/s	—
27	DeepSeek V4 Pro (Reasoning, High Effort)DeepSeek	49.8	43.2	—	46 tok/s	View →
28	GLM-5 (Reasoning)Z.ai	49.8	44.2	—	79 tok/s	View →
29	Claude Opus 4.5 (Reasoning)Anthropic	49.7	47.8	91.3	54 tok/s	View →
30	MiniMax-M2.7MiniMax	49.6	41.9	—	67 tok/s	View →
31	Grok 4.20 0309 v2 (Reasoning)xAI	49.3	40.5	—	172 tok/s	—
32	MiMo-V2-ProXiaomi	49.2	41.4	—	43 tok/s	—
33	MiMo-V2.5Xiaomi	49.0	42.1	—	82 tok/s	—
34	GPT-5.2 Codex (xhigh)OpenAI	49.0	43.0	—	108 tok/s	View →
35	GPT-5.4 mini (xhigh)OpenAI	48.9	51.5	—	157 tok/s	View →
36	Grok 4.3 (medium)xAI	48.8	35.1	—	127 tok/s	View →
37	Grok 4.20 0309 (Reasoning)xAI	48.5	42.2	—	178 tok/s	—
38	Gemini 3 Pro Preview (high)Google	48.4	46.5	95.7	—	—
39	GPT-5.4 (low)OpenAI	47.9	45.6	—	64 tok/s	View →
40	Nemotron 3 Ultra 550B A55B (Reasoning)NVIDIA	47.7	37.6	—	390 tok/s	—
41	GPT-5.1 (high)OpenAI	47.7	44.7	94.0	125 tok/s	View →
42	GLM-5-TurboZ.ai	46.8	36.8	—	—	View →
43	Kimi K2.5 (Reasoning)Moonshot	46.8	39.6	—	34 tok/s	View →
44	GPT-5.2 (medium)OpenAI	46.6	44.2	96.7	—	View →
45	DeepSeek V4 Flash (Reasoning, Max Effort)DeepSeek	46.5	38.7	—	120 tok/s	View →
46	Claude Opus 4.6 (Non-reasoning, High Effort)Anthropic	46.5	47.6	—	44 tok/s	View →
47	Gemini 3 Flash Preview (Reasoning)Google	46.4	42.6	97.0	181 tok/s	View →
48	DeepSeek V4 Flash (Reasoning, High Effort)DeepSeek	46.0	39.8	—	—	View →
49	Qwen3.6 27B (Reasoning)Alibaba	45.8	36.5	—	56 tok/s	View →
50	Qwen3.5 397B A17B (Reasoning)Alibaba	45.0	41.3	—	52 tok/s	View →
51	MiMo-V2-Omni-0327Xiaomi	44.9	36.9	—	88 tok/s	—
52	GPT-5 (high)OpenAI	44.6	36.0	94.3	119 tok/s	View →
53	GPT-5 Codex (high)OpenAI	44.6	38.9	98.7	174 tok/s	View →
54	Claude Sonnet 4.6 (Non-reasoning, High Effort)Anthropic	44.4	46.4	—	46 tok/s	View →
55	GPT-5.4 nano (xhigh)OpenAI	44.0	43.9	—	161 tok/s	View →
56	Grok 4.3 (low)xAI	43.9	31.6	—	146 tok/s	View →
57	KAT Coder Pro V2KwaiKAT	43.8	45.6	—	119 tok/s	—
58	GLM-5.1 (Non-reasoning)Z.ai	43.8	35.8	—	50 tok/s	View →
59	Qwen3.6 35B A3B (Reasoning)Alibaba	43.5	35.2	—	162 tok/s	View →
60	MiMo-V2-OmniXiaomi	43.4	35.5	—	84 tok/s	—
61	Gemini 3.5 Flash (minimal)Google	43.3	47.1	—	192 tok/s	—
62	GPT-5.1 Codex (high)OpenAI	43.1	36.6	95.7	182 tok/s	View →
63	Claude Opus 4.5 (Non-reasoning)Anthropic	43.1	42.9	62.7	48 tok/s	View →
64	Claude 4.5 Sonnet (Reasoning)Anthropic	43.0	38.6	88.0	53 tok/s	View →
65	Kimi K2.6 (Non-reasoning)Moonshot	42.9	38.4	—	41 tok/s	View →
66	GLM 5V Turbo (Reasoning)Z.ai	42.9	36.2	—	—	View →
67	Claude Sonnet 4.6 (Non-reasoning, Low Effort)Anthropic	42.6	43.0	—	46 tok/s	View →
68	Step 3.7 FlashStepFun	42.6	37.1	—	392 tok/s	—
69	GLM-4.7 (Reasoning)Z.ai	42.1	36.3	95.0	80 tok/s	View →
70	Qwen3.5 27B (Reasoning)Alibaba	42.1	34.9	—	83 tok/s	View →
71	GPT-5 (medium)OpenAI	42.0	38.9	91.7	84 tok/s	View →
72	Claude 4.1 Opus (Reasoning)Anthropic	42.0	36.5	80.3	34 tok/s	View →
73	Hy3-preview (Reasoning)Tencent	41.9	36.5	—	94 tok/s	—
74	MiniMax-M2.5MiniMax	41.9	37.4	—	202 tok/s	View →
75	GPT-5.5 Instant (May 2026)OpenAI	41.8	45.1	—	—	—
76	DeepSeek V3.2 (Reasoning)DeepSeek	41.7	36.7	92.0	—	View →
77	Qwen3.5 122B A10B (Reasoning)Alibaba	41.6	34.7	—	140 tok/s	View →
78	MiMo-V2-Flash (Feb 2026)Xiaomi	41.5	33.5	—	125 tok/s	—
79	Grok 4xAI	41.5	40.5	92.7	—	View →
80	Gemini 3 Pro Preview (low)Google	41.3	39.4	86.7	—	—
81	GPT-5 mini (high)OpenAI	41.2	35.3	90.7	90 tok/s	View →
82	GPT-5.5 (Non-reasoning)OpenAI	40.9	48.6	—	51 tok/s	View →
83	Kimi K2 ThinkingMoonshot	40.9	34.8	94.7	131 tok/s	View →
84	o3-proOpenAI	40.7	—	—	24 tok/s	View →
85	GLM-5 (Non-reasoning)Z.ai	40.6	39.0	—	65 tok/s	View →
86	Qwen3.5 397B A17B (Non-reasoning)Alibaba	40.1	37.4	—	53 tok/s	View →
87	Qwen3 Max ThinkingAlibaba	39.8	30.5	—	—	View →
88	MiniMax-M2.1MiniMax	39.4	32.8	82.7	175 tok/s	View →
89	DeepSeek V4 Pro (Non-reasoning)DeepSeek	39.3	38.4	—	45 tok/s	View →
90	Gemma 4 31B (Reasoning)Google	39.2	38.7	—	35 tok/s	View →
91	Mistral Medium 3.5Mistral	39.2	35.4	—	139 tok/s	—
92	GPT-5 (low)OpenAI	39.2	30.7	83.0	86 tok/s	View →
93	MiMo-V2-Flash (Reasoning)Xiaomi	39.2	31.8	96.3	130 tok/s	—
94	Claude 4 Opus (Reasoning)Anthropic	39.0	34.0	73.3	37 tok/s	View →
95	GPT-5 mini (medium)OpenAI	38.9	32.8	85.0	88 tok/s	View →
96	Claude 4 Sonnet (Reasoning)Anthropic	38.7	34.1	74.3	46 tok/s	View →
97	Qwen3.5 Omni PlusAlibaba	38.6	27.6	—	53 tok/s	—
98	GPT-5.1 Codex mini (high)OpenAI	38.6	36.4	91.7	211 tok/s	View →
99	Grok 4.1 Fast (Reasoning)xAI	38.6	30.9	89.3	—	View →
100	Step 3.5 Flash 2603StepFun	38.5	34.6	—	229 tok/s	—
101	Ring-2.6-1TInclusionAI	38.5	33.3	—	125 tok/s	—
102	o3OpenAI	38.4	38.4	88.3	113 tok/s	View →
103	GPT-5.4 nano (medium)OpenAI	38.1	35.0	—	160 tok/s	View →
104	Step 3.5 FlashStepFun	37.8	31.6	—	220 tok/s	—
105	GPT-5.4 mini (medium)OpenAI	37.7	37.5	—	166 tok/s	View →
106	Kimi K2.5 (Non-reasoning)Moonshot	37.3	25.8	—	33 tok/s	View →
107	Command A+Cohere	37.2	29.3	—	211 tok/s	—
108	Qwen3.5 27B (Non-reasoning)Alibaba	37.2	33.4	—	91 tok/s	View →
109	Claude 4.5 Haiku (Reasoning)Anthropic	37.1	32.6	83.7	135 tok/s	View →
110	Qwen3.6 27B (Non-reasoning)Alibaba	37.1	26.6	—	53 tok/s	View →
111	Claude 4.5 Sonnet (Non-reasoning)Anthropic	37.1	33.5	37.0	43 tok/s	View →
112	Qwen3.5 35B A3B (Reasoning)Alibaba	37.1	30.3	—	133 tok/s	View →
113	DeepSeek V4 Flash (Non-reasoning)DeepSeek	36.5	35.2	—	109 tok/s	View →
114	JT-35B-FlashChina Mobile	36.1	28.9	—	—	—
115	MiniMax-M2MiniMax	36.1	29.2	78.3	112 tok/s	View →
116	NVIDIA Nemotron 3 Super 120B A12B (Reasoning)NVIDIA	36.0	31.2	—	150 tok/s	View →
117	KAT-Coder-Pro V1KwaiKAT	36.0	18.3	94.7	116 tok/s	—
118	Claude 4.1 Opus (Non-reasoning)Anthropic	36.0	—	—	33 tok/s	View →
119	Qwen3.5 122B A10B (Non-reasoning)Alibaba	35.9	31.6	—	152 tok/s	View →
120	Nova 2.0 Pro Preview (medium)Amazon	35.7	30.4	89.0	125 tok/s	—
121	MiMo-V2.5-Pro (Non-reasoning)Xiaomi	35.6	36.8	—	46 tok/s	—
122	GPT-5.4 (Non-reasoning)OpenAI	35.4	41.0	—	59 tok/s	View →
123	Grok 4 Fast (Reasoning)xAI	35.1	27.4	89.7	—	—
124	Gemini 3 Flash Preview (Non-reasoning)Google	35.0	37.8	55.7	191 tok/s	View →
125	Claude 3.7 Sonnet (Reasoning)Anthropic	34.7	27.6	56.3	—	—
126	Gemini 2.5 ProGoogle	34.6	32.0	87.7	125 tok/s	View →
127	Nova 2.0 Lite (high)Amazon	34.5	23.4	94.3	154 tok/s	—
128	GLM-4.7 (Non-reasoning)Z.ai	34.2	32.0	48.0	81 tok/s	View →
129	DeepSeek V3.1 Terminus (Reasoning)DeepSeek	33.9	33.7	89.7	—	View →
130	Hy3-preview (Non-reasoning)Tencent	33.7	34.3	—	84 tok/s	—
131	Ling-2.6-1TInclusionAI	33.6	33.1	—	—	—
132	GPT-5.2 (Non-reasoning)OpenAI	33.6	34.7	51.0	61 tok/s	View →
133	Gemini 3.1 Flash-LiteGoogle	33.5	30.1	—	291 tok/s	View →
134	Doubao Seed CodeByteDance Seed	33.5	31.3	79.3	—	—
135	gpt-oss-120b (high)OpenAI	33.3	28.6	93.4	360 tok/s	View →
136	o4-mini (high)OpenAI	33.1	25.6	90.7	153 tok/s	View →
137	Claude 4 Sonnet (Non-reasoning)Anthropic	33.0	30.6	38.0	45 tok/s	View →
138	Claude 4 Opus (Non-reasoning)Anthropic	33.0	—	36.3	35 tok/s	View →
139	DeepSeek V3.2 Exp (Reasoning)DeepSeek	32.9	33.3	87.7	—	View →
140	Mercury 2Inception	32.8	30.6	—	961 tok/s	—
141	GLM-4.6 (Reasoning)Z.ai	32.5	29.5	86.0	40 tok/s	View →
142	Qwen3 Max Thinking (Preview)Alibaba	32.5	24.5	82.3	52 tok/s	—
143	Qwen3.5 9B (Reasoning)Alibaba	32.4	25.3	—	73 tok/s	View →
144	Gemma 4 31B (Non-reasoning)Google	32.3	33.9	—	28 tok/s	View →
145	K-EXAONE (Reasoning)LG AI Research	32.1	27.0	90.3	—	—
146	DeepSeek V3.2 (Non-reasoning)DeepSeek	32.1	34.6	59.0	—	View →
147	Grok 3 mini Reasoning (high)xAI	32.1	25.2	84.7	58 tok/s	View →
148	Nova 2.0 Pro Preview (low)Amazon	31.9	24.5	63.3	133 tok/s	—
149	Trinity Large ThinkingArcee	31.9	27.2	—	171 tok/s	View →
150	Qwen3.6 35B A3B (Non-reasoning)Alibaba	31.5	17.6	—	158 tok/s	View →
151	Qwen3 MaxAlibaba	31.4	26.4	80.7	49 tok/s	View →
152	Gemma 4 26B A4B (Reasoning)Google	31.2	22.4	—	—	View →
153	Gemini 2.5 Flash Preview (Sep '25) (Reasoning)Google	31.1	24.6	78.3	—	View →
154	Claude 4.5 Haiku (Non-reasoning)Anthropic	31.0	29.6	39.0	98 tok/s	View →
155	Grok 4.3 (Non-reasoning)xAI	31.0	25.1	—	127 tok/s	View →
156	Kimi K2 0905Moonshot	30.9	25.9	57.3	24 tok/s	View →
157	Claude 3.7 Sonnet (Non-reasoning)Anthropic	30.8	26.7	21.0	—	—
158	Qwen3.5 35B A3B (Non-reasoning)Alibaba	30.7	16.8	—	147 tok/s	View →
159	o1OpenAI	30.7	20.5	—	120 tok/s	View →
160	MiMo-V2-Flash (Non-reasoning)Xiaomi	30.3	25.8	67.7	122 tok/s	—
161	Gemini 2.5 Pro Preview (Mar' 25)Google	30.3	46.7	—	—	View →
162	EXAONE 4.5 33BLG AI Research	30.2	23.0	—	—	—
163	GLM-4.6 (Non-reasoning)Z.ai	30.2	30.2	44.3	48 tok/s	View →
164	GLM-4.7-Flash (Reasoning)Z.ai	30.1	25.9	—	93 tok/s	View →
165	Nova 2.0 Lite (medium)Amazon	29.7	23.9	88.7	136 tok/s	—
166	Grok 4.20 0309 (Non-reasoning)xAI	29.7	25.4	—	164 tok/s	View →
167	Gemini 2.5 Pro Preview (May' 25)Google	29.5	—	—	—	View →
168	Qwen3 235B A22B 2507 (Reasoning)Alibaba	29.5	23.2	91.0	62 tok/s	View →
169	DeepSeek V3.2 SpecialeDeepSeek	29.4	37.9	96.7	—	—
170	ERNIE 5.0 Thinking PreviewBaidu	29.1	29.2	85.0	—	—
171	Grok 4.20 0309 v2 (Non-reasoning)xAI	29.0	22.0	—	164 tok/s	View →
172	Grok Code Fast 1xAI	28.7	23.7	43.3	—	—
173	DeepSeek V3.1 Terminus (Non-reasoning)DeepSeek	28.5	31.9	53.7	—	View →
174	Nemotron Cascade 2 30B A3BNVIDIA	28.4	25.8	—	—	—
175	DeepSeek V3.2 Exp (Non-reasoning)DeepSeek	28.4	30.0	57.7	—	View →
176	Qwen3 Coder NextAlibaba	28.3	22.9	—	102 tok/s	View →
177	Apriel-v1.5-15B-ThinkerServiceNow	28.3	18.7	87.5	—	—
178	DeepSeek V3.1 (Non-reasoning)DeepSeek	28.1	28.4	49.7	—	View →
179	Nova 2.0 Omni (medium)Amazon	28.0	15.1	89.7	—	—
180	Mistral Small 4 (Reasoning)Mistral	27.8	24.3	—	161 tok/s	—
181	DeepSeek V3.1 (Reasoning)DeepSeek	27.7	29.7	89.7	—	View →
182	Apriel-v1.6-15B-ThinkerServiceNow	27.6	22.0	88.0	—	—
183	Qwen3 VL 235B A22B (Reasoning)Alibaba	27.6	20.9	88.3	33 tok/s	—
184	GPT-5.1 (Non-reasoning)OpenAI	27.4	27.3	38.0	107 tok/s	View →
185	Qwen3.5 9B (Non-reasoning)Alibaba	27.3	21.3	—	—	View →
186	Gemma 4 26B A4B (Non-reasoning)Google	27.1	29.1	—	62 tok/s	View →
187	Magistral Medium 1.2Mistral	27.1	21.7	82.0	39 tok/s	View →
188	Qwen3.5 4B (Reasoning)Alibaba	27.1	17.5	—	198 tok/s	View →
189	DeepSeek R1 0528 (May '25)DeepSeek	27.1	24.0	76.0	—	View →
190	Gemini 2.5 Flash (Reasoning)Google	27.0	22.2	73.3	219 tok/s	View →
191	GPT-5 nano (high)OpenAI	26.8	20.3	83.7	153 tok/s	View →
192	Qwen3 Next 80B A3B (Reasoning)Alibaba	26.7	19.5	84.3	142 tok/s	View →
193	GLM-4.5 (Reasoning)Z AI	26.4	26.3	73.7	50 tok/s	—
194	GPT-4.1OpenAI	26.3	21.8	34.7	132 tok/s	View →
195	Kimi K2Kimi	26.3	22.1	57.0	25 tok/s	—
196	Ling 2.6 FlashInclusionAI	26.2	23.2	—	—	—
197	Qwen3 Max (Preview)Alibaba	26.1	25.5	75.0	48 tok/s	—
198	Solar Pro 3Upstage	25.9	13.3	—	—	—
199	Qwen3.5 Omni FlashAlibaba	25.9	14.0	—	226 tok/s	—
200	o3-miniOpenAI	25.9	17.9	—	200 tok/s	View →
201	GPT-5 nano (medium)OpenAI	25.9	22.9	78.3	168 tok/s	View →
202	o1-proOpenAI	25.8	—	—	—	View →
203	Gemini 2.5 Flash Preview (Sep '25) (Non-reasoning)Google	25.7	22.1	56.7	—	View →
204	JT-MINIChina Mobile	25.4	21.2	—	—	—
205	o3-mini (high)OpenAI	25.2	17.3	—	212 tok/s	View →
206	Grok 3xAI	25.2	19.8	58.0	—	—
207	Seed-OSS-36B-InstructByteDance Seed	25.2	16.7	84.7	39 tok/s	—
208	Qwen3 235B A22B 2507 InstructAlibaba	25.0	22.1	71.7	44 tok/s	View →
209	Qwen3 Coder 480B A35B InstructAlibaba	24.8	24.6	39.3	68 tok/s	—
210	Qwen3 VL 32B (Reasoning)Alibaba	24.7	14.5	84.7	99 tok/s	—
211	Nova 2.0 Lite (low)Amazon	24.6	13.6	46.7	165 tok/s	—
212	Sonar Reasoning ProPerplexity	24.6	—	—	—	View →
213	gpt-oss-120b (low)OpenAI	24.5	15.5	66.7	364 tok/s	View →
214	gpt-oss-20B (high)OpenAI	24.5	18.5	89.3	268 tok/s	View →
215	GPT-5.4 nano (Non-Reasoning)OpenAI	24.4	27.9	—	155 tok/s	View →
216	MiniMax M1 80kMiniMax	24.4	14.5	61.0	—	View →
217	NVIDIA Nemotron 3 Nano 30B A3B (Reasoning)NVIDIA	24.3	19.0	91.0	148 tok/s	View →
218	Gemini 2.5 Flash Preview (Reasoning)Google	24.3	—	—	—	View →
219	K2 Think V2MBZUAI Institute of Foundation Models	24.1	15.5	—	—	—
220	LongCat Flash LiteLongCat	23.9	16.5	—	—	—
221	GPT-5 (minimal)OpenAI	23.9	25.0	31.7	71 tok/s	—
222	HyperCLOVA X SEED Think (32B)Naver	23.7	17.5	59.0	—	—
223	o1-previewOpenAI	23.7	34.0	—	—	—
224	Grok 4.1 Fast (Non-reasoning)xAI	23.6	19.5	34.3	—	—
225	K-EXAONE (Non-reasoning)LG AI Research	23.4	13.5	44.0	—	—
226	GLM-4.6V (Reasoning)Z.ai	23.4	19.7	85.3	61 tok/s	View →
227	GPT-5.4 mini (Non-Reasoning)OpenAI	23.3	25.3	—	150 tok/s	View →
228	Nova 2.0 Omni (low)Amazon	23.2	13.9	56.0	—	—
229	GLM-4.5-AirZ.ai	23.2	23.8	80.7	68 tok/s	View →
230	Nova 2.0 Pro Preview (Non-reasoning)Amazon	23.1	20.5	30.7	134 tok/s	—
231	Mi:dm K 2.5 ProKorea Telecom	23.1	12.6	76.7	—	—
232	Grok 4 Fast (Non-reasoning)xAI	23.1	19.0	41.3	—	—
233	GPT-4.1 miniOpenAI	22.9	18.5	46.3	85 tok/s	View →
234	Mistral Large 3Mistral	22.8	22.7	38.0	60 tok/s	View →
235	Ring-1TInclusionAI	22.8	16.8	89.3	—	—
236	Qwen3.5 4B (Non-reasoning)Alibaba	22.6	13.7	—	206 tok/s	View →
237	Qwen3 30B A3B 2507 (Reasoning)Alibaba	22.4	14.6	56.3	140 tok/s	—
238	DeepSeek V3 0324DeepSeek	22.3	22.0	41.0	—	View →
239	INTELLECT-3PrimeIntellect	22.2	19.1	88.0	—	View →
240	GLM-4.7-Flash (Non-reasoning)Z.ai	22.1	11.0	—	120 tok/s	View →
241	Devstral 2Mistral	22.0	23.7	36.7	38 tok/s	—
242	GPT-5 (ChatGPT)OpenAI	21.8	21.2	48.3	167 tok/s	—
243	Solar Open 100B (Reasoning)Upstage	21.7	10.5	—	—	—
244	Gemini 2.5 Flash-Lite Preview (Sep '25) (Reasoning)Google	21.6	18.2	68.7	—	View →
245	Grok 3 Reasoning BetaxAI	21.6	—	—	—	—
246	Nemotron 3 Nano Omni 30B A3B ReasoningNVIDIA	21.4	14.8	—	294 tok/s	—
247	Mistral Medium 3.1Mistral	21.3	18.3	38.3	69 tok/s	—
248	MiniMax M1 40kMiniMax	20.9	14.1	13.7	—	—
249	gpt-oss-20B (low)OpenAI	20.8	14.4	62.3	274 tok/s	View →
250	Qwen3 VL 235B A22B InstructAlibaba	20.8	16.5	70.7	48 tok/s	View →
251	GPT-5 mini (minimal)OpenAI	20.7	21.9	46.7	89 tok/s	—
252	K2-V2 (high)MBZUAI Institute of Foundation Models	20.6	16.1	78.3	—	—
253	Gemini 2.5 Flash (Non-reasoning)Google	20.6	17.8	60.3	186 tok/s	View →
254	o1-miniOpenAI	20.4	—	—	—	—
255	Qwen3 Next 80B A3B InstructAlibaba	20.1	15.3	66.3	132 tok/s	View →
256	Tri-21B-think PreviewTrillion Labs	20.0	7.4	—	—	—
257	GPT-4.5 (Preview)OpenAI	20.0	—	—	—	—
258	Qwen3 Coder 30B A3B InstructAlibaba	20.0	19.4	29.0	82 tok/s	View →
259	Qwen3 235B A22B (Reasoning)Alibaba	19.8	17.4	82.0	60 tok/s	View →
260	Qwen3 VL 30B A3B (Reasoning)Alibaba	19.7	13.1	82.3	127 tok/s	View →
261	QwQ 32BAlibaba	19.7	—	29.0	31 tok/s	—
262	Gemini 2.0 Flash Thinking Experimental (Jan '25)Google	19.6	24.1	—	—	—
263	Devstral Small 2Mistral	19.5	20.7	34.3	51 tok/s	—
264	Gemini 2.5 Flash-Lite Preview (Sep '25) (Non-reasoning)Google	19.4	14.5	46.7	—	View →
265	Motif-2-12.7B-ReasoningMotif Technologies	19.1	11.9	80.3	—	—
266	Nova PremierAmazon	19.0	13.8	17.3	64 tok/s	—
267	Ling-1TInclusionAI	19.0	18.8	71.3	—	—
268	Gemma 4 E4B (Reasoning)Google	18.8	13.7	—	—	—
269	Mistral Medium 3Mistral	18.8	13.6	30.3	42 tok/s	—
270	Magistral Medium 1Mistral	18.8	16.0	40.3	—	—
271	DeepSeek R1 (Jan '25)DeepSeek	18.8	15.9	68.0	—	—
272	Solar Pro 2 (Preview) (Reasoning)Upstage	18.8	—	—	—	—
273	Llama Nemotron Super 49B v1.5 (Reasoning)NVIDIA	18.7	15.1	76.7	44 tok/s	—
274	K2-V2 (medium)MBZUAI Institute of Foundation Models	18.7	14.0	64.7	—	—
275	Claude 3.5 HaikuAnthropic	18.7	10.7	—	—	—
276	Devstral MediumMistral	18.7	15.9	4.7	29 tok/s	—
277	Mistral Small 4 (Non-reasoning)Mistral	18.6	16.4	—	154 tok/s	—
278	Hermes 4 - Llama-3.1 405B (Reasoning)Nous Research	18.6	16.0	69.7	39 tok/s	—
279	Tri-21B-ThinkTrillion Labs	18.6	6.3	—	—	—
280	GPT-4o (Aug '24)OpenAI	18.6	16.6	—	90 tok/s	View →
281	GPT-4o (March 2025, chatgpt-4o-latest)OpenAI	18.6	—	25.7	—	—
282	Gemini 2.0 Flash (Feb '25)Google	18.5	13.6	21.7	—	View →
283	Llama 3.3 Nemotron Super 49B v1 (Reasoning)NVIDIA	18.5	9.4	54.7	—	—
284	Llama 4 MaverickMeta	18.4	15.6	19.3	114 tok/s	—
285	Magistral Small 1.2Mistral	18.2	14.8	80.3	110 tok/s	View →
286	Sarvam 105B (high)Sarvam	18.2	9.8	—	94 tok/s	—
287	MiniCPM5-1B (Reasoning)OpenBMB	18.2	1.5	—	—	—
288	Qwen3 4B 2507 (Reasoning)Alibaba	18.2	9.5	82.7	—	—
289	Gemini 2.0 Pro Experimental (Feb '25)Google	18.1	25.5	—	—	—
290	Nova 2.0 Lite (Non-reasoning)Amazon	18.0	12.5	33.7	209 tok/s	—
291	Claude 3 OpusAnthropic	18.0	19.5	—	—	—
292	Devstral Small (May '25)Mistral	18.0	12.2	—	—	—
293	MiniCPM5-1B (Non-reasoning)OpenBMB	17.9	0.5	—	—	—
294	Sonar ReasoningPerplexity	17.9	—	—	—	View →
295	Gemini 2.5 Flash Preview (Non-reasoning)Google	17.8	—	—	—	View →
296	Hermes 4 - Llama-3.1 405B (Non-reasoning)Nous Research	17.6	18.1	15.3	42 tok/s	—
297	Gemini 2.5 Flash-Lite (Reasoning)Google	17.6	9.5	53.3	307 tok/s	View →
298	Llama 3.1 Instruct 405BMeta	17.4	14.5	3.0	54 tok/s	—
299	GPT-4o (Nov '24)OpenAI	17.3	16.7	6.0	143 tok/s	View →
300	DeepSeek R1 Distill Qwen 32BDeepSeek	17.2	—	63.0	—	View →
301	Qwen3 VL 32B InstructAlibaba	17.2	15.6	68.3	76 tok/s	—
302	GLM-4.6V (Non-reasoning)Z.ai	17.1	11.1	26.3	68 tok/s	View →
303	Qwen3 235B A22B (Non-reasoning)Alibaba	17.0	14.0	23.7	65 tok/s	View →
304	Gemini 2.0 Flash (experimental)Google	16.8	—	—	—	—
305	Magistral Small 1Mistral	16.8	11.1	41.3	—	View →
306	EXAONE 4.0 32B (Reasoning)LG AI Research	16.7	14.0	80.0	—	—
307	Qwen3 VL 8B (Reasoning)Alibaba	16.7	9.8	30.7	137 tok/s	—
308	Nova 2.0 Omni (Non-reasoning)Amazon	16.6	13.8	37.0	—	—
309	DeepSeek V3 (Dec '24)DeepSeek	16.5	16.4	26.0	—	View →
310	Qwen3 32B (Reasoning)Alibaba	16.5	13.8	73.0	98 tok/s	View →
311	DeepSeek R1 0528 Qwen3 8BDeepSeek	16.4	7.8	63.7	—	—
312	Qwen3.5 2B (Reasoning)Alibaba	16.3	3.5	—	—	View →
313	Qwen2.5 MaxAlibaba	16.3	—	—	—	—
314	Qwen3 14B (Reasoning)Alibaba	16.2	13.1	55.7	63 tok/s	—
315	Nanbeige4.1-3BNanbeige	16.1	8.9	—	—	—
316	Ministral 3 14BMistral	16.0	10.9	30.0	88 tok/s	View →
317	Hermes 4 - Llama-3.1 70B (Reasoning)Nous Research	16.0	14.4	68.7	88 tok/s	—
318	Gemini 1.5 Pro (Sep '24)Google	16.0	23.6	—	—	—
319	DeepSeek R1 Distill Llama 70BDeepSeek	16.0	11.4	53.7	45 tok/s	View →
320	Solar Pro 2 (Preview) (Non-reasoning)Upstage	16.0	—	—	—	—
321	Qwen3 VL 30B A3B InstructAlibaba	16.0	14.3	72.3	125 tok/s	View →
322	Claude 3.5 Sonnet (Oct '24)Anthropic	15.9	30.2	—	—	—
323	Falcon-H1R-7BTII UAE	15.8	9.8	80.0	—	—
324	DeepSeek R1 Distill Qwen 14BDeepSeek	15.8	—	55.7	—	View →
325	Ling-flash-2.0InclusionAI	15.7	16.7	65.3	72 tok/s	—
326	Qwen3 Omni 30B A3B (Reasoning)Alibaba	15.6	12.7	74.0	86 tok/s	—
327	Qwen2.5 Instruct 72BAlibaba	15.6	11.9	14.0	—	View →
328	Step3 VL 10BStepFun	15.5	13.9	—	—	—
329	SonarPerplexity	15.5	—	—	—	View →
330	Qwen3 30B A3B (Reasoning)Alibaba	15.3	11.0	72.3	67 tok/s	View →
331	Gemma 4 E2B (Reasoning)Google	15.2	9.0	—	—	—
332	Devstral Small (Jul '25)Mistral	15.2	12.1	29.3	36 tok/s	—
333	Sonar ProPerplexity	15.2	—	—	—	View →
334	QwQ 32B-PreviewAlibaba	15.2	—	—	—	—
335	Mistral Large 2 (Nov '24)Mistral	15.1	13.8	14.0	59 tok/s	—
336	Mistral Small 3.2Mistral	15.1	13.3	27.0	127 tok/s	—
337	GLM-4.5V (Reasoning)Z.ai	15.1	10.9	73.0	31 tok/s	View →
338	Llama 3.1 Nemotron Ultra 253B v1 (Reasoning)NVIDIA	15.0	13.1	63.7	53 tok/s	—
339	ERNIE 4.5 300B A47BBaidu	15.0	14.5	41.3	24 tok/s	—
340	Qwen3 30B A3B 2507 InstructAlibaba	15.0	14.2	66.3	114 tok/s	—
341	Solar Pro 2 (Reasoning)Upstage	14.9	12.1	61.3	—	—
342	NVIDIA Nemotron Nano 12B v2 VL (Reasoning)NVIDIA	14.9	11.7	75.0	191 tok/s	View →
343	Gemma 4 E4B (Non-reasoning)Google	14.8	6.4	—	—	—
344	Ministral 3 8BMistral	14.8	10.0	31.7	101 tok/s	View →
345	NVIDIA Nemotron Nano 9B V2 (Reasoning)NVIDIA	14.8	8.3	69.7	115 tok/s	View →
346	NVIDIA Nemotron 3 Nano 4BNVIDIA	14.7	10.0	—	—	—
347	Granite 4.1 30BIBM	14.7	10.1	—	—	—
348	Qwen3.5 2B (Non-reasoning)Alibaba	14.7	4.9	—	315 tok/s	View →
349	Gemini 2.0 Flash-Lite (Feb '25)Google	14.7	—	—	—	View →
350	Llama Nemotron Super 49B v1.5 (Non-reasoning)NVIDIA	14.6	10.5	8.0	44 tok/s	—
351	Llama 3.3 Instruct 70BMeta	14.5	10.7	7.7	91 tok/s	View →
352	GPT-4o (May '24)OpenAI	14.5	24.2	—	94 tok/s	View →
353	Gemini 2.0 Flash-Lite (Preview)Google	14.5	—	—	—	—
354	Mistral Small 3.1Mistral	14.5	13.9	3.7	155 tok/s	View →
355	Qwen3 32B (Non-reasoning)Alibaba	14.5	—	19.7	97 tok/s	View →
356	Kimi Linear 48B A3B InstructKimi	14.4	14.2	36.3	—	—
357	K2-V2 (low)MBZUAI Institute of Foundation Models	14.4	10.5	35.3	—	—
358	Llama 3.1 Nemotron Nano 4B v1.1 (Reasoning)NVIDIA	14.4	—	50.0	—	—
359	Llama 3.3 Nemotron Super 49B v1 (Non-reasoning)NVIDIA	14.3	7.6	7.7	—	—
360	Qwen3 VL 8B InstructAlibaba	14.3	7.3	27.3	148 tok/s	View →
361	Claude 3.5 Sonnet (June '24)Anthropic	14.2	26.0	—	—	—
362	Qwen3 4B (Reasoning)Alibaba	14.2	—	22.3	—	—
363	GPT-4o (ChatGPT)OpenAI	14.1	—	—	—	—
364	Llama 3.1 Tulu3 405BAllen Institute for AI	14.1	—	—	—	—
365	Ring-flash-2.0InclusionAI	14.0	10.6	83.7	—	—
366	Pixtral LargeMistral	14.0	—	2.3	58 tok/s	—
367	Olmo 3.1 32B ThinkAllen Institute for AI	13.9	9.8	77.3	—	—
368	Grok 2 (Dec '24)xAI	13.9	—	—	—	—
369	GPT-5 nano (minimal)OpenAI	13.8	14.2	27.3	154 tok/s	—
370	Gemini 1.5 Flash (Sep '24)Google	13.8	—	—	—	—
371	GPT-4 TurboOpenAI	13.7	21.5	—	26 tok/s	—
372	Qwen3 VL 4B (Reasoning)Alibaba	13.7	6.7	25.7	—	—
373	Solar Pro 2 (Non-reasoning)Upstage	13.6	11.3	30.0	—	—
374	Llama 4 ScoutMeta	13.5	6.7	14.0	106 tok/s	View →
375	Command ACohere	13.5	9.9	13.0	68 tok/s	—
376	Nova ProAmazon	13.5	11.0	7.0	—	View →
377	Llama 3.1 Nemotron Instruct 70BNVIDIA	13.4	10.8	11.0	281 tok/s	—
378	Grok BetaxAI	13.3	—	—	—	—
379	NVIDIA Nemotron Nano 9B V2 (Non-reasoning)NVIDIA	13.2	7.5	62.3	134 tok/s	View →
380	NVIDIA Nemotron 3 Nano 30B A3B (Non-reasoning)NVIDIA	13.2	15.8	13.3	89 tok/s	View →
381	Qwen2.5 Instruct 32BAlibaba	13.2	—	—	—	—
382	Qwen3 8B (Reasoning)Alibaba	13.2	9.0	19.0	63 tok/s	—
383	GPT-4.1 nanoOpenAI	13.0	11.2	24.0	107 tok/s	View →
384	Mistral Large 2 (Jul '24)Mistral	13.0	—	0.0	—	—
385	Qwen2.5 Coder Instruct 32BAlibaba	12.9	—	—	—	—
386	Qwen3 4B 2507 InstructAlibaba	12.9	9.0	52.3	—	—
387	GPT-4OpenAI	12.8	13.1	—	39 tok/s	—
388	Qwen3 14B (Non-reasoning)Alibaba	12.8	12.4	58.0	65 tok/s	—
389	MiniCPM-V 4.6 1.3BOpenBMB	12.7	0.7	—	—	—
390	Gemini 2.5 Flash-Lite (Non-reasoning)Google	12.7	7.4	35.3	231 tok/s	View →
391	Mistral Small 3Mistral	12.7	—	4.3	151 tok/s	—
392	Nova LiteAmazon	12.7	5.1	7.0	185 tok/s	View →
393	GLM-4.5V (Non-reasoning)Z.ai	12.7	10.8	15.3	48 tok/s	View →
394	Hermes 4 - Llama-3.1 70B (Non-reasoning)Nous Research	12.6	9.2	11.3	84 tok/s	—
395	GPT-4o miniOpenAI	12.6	—	14.7	68 tok/s	View →
396	Llama 3.1 Instruct 70BMeta	12.5	10.9	4.0	43 tok/s	—
397	DeepSeek-V2.5 (Dec '24)DeepSeek	12.5	—	—	—	—
398	Qwen3 4B (Non-reasoning)Alibaba	12.5	—	—	—	—
399	Qwen3 30B A3B (Non-reasoning)Alibaba	12.5	13.3	21.7	69 tok/s	View →
400	Granite 4.1 8BIBM	12.4	7.3	—	131 tok/s	—
401	Sarvam 30B (high)Sarvam	12.3	7.9	—	167 tok/s	—
402	Gemini 2.0 Flash Thinking Experimental (Dec '24)Google	12.3	—	—	—	—
403	Claude 3 HaikuAnthropic	12.3	6.7	—	—	View →
404	DeepSeek-V2.5DeepSeek	12.3	—	—	—	—
405	Olmo 3.1 32B InstructAllen Institute for AI	12.2	5.6	—	—	—
406	Gemma 4 E2B (Non-reasoning)Google	12.1	8.3	—	—	—
407	Mistral SabaMistral	12.1	—	—	—	—
408	DeepSeek R1 Distill Llama 8BDeepSeek	12.1	—	41.3	—	—
409	Olmo 3 32B ThinkAllen Institute for AI	12.1	10.5	73.7	—	—
410	R1 1776Perplexity	12.0	—	—	—	—
411	Gemini 1.5 Pro (May '24)Google	12.0	19.8	—	—	—
412	Reka Flash (Sep '24)Reka AI	12.0	—	—	86 tok/s	—
413	Qwen2.5 TurboAlibaba	12.0	—	—	66 tok/s	—
414	Llama 3.2 Instruct 90B (Vision)Meta	11.9	—	—	46 tok/s	—
415	Solar MiniUpstage	11.9	—	—	76 tok/s	—
416	Llama 3.1 Instruct 8BMeta	11.8	4.9	4.3	203 tok/s	—
417	Grok-1xAI	11.7	—	—	—	—
418	EXAONE 4.0 32B (Non-reasoning)LG AI Research	11.7	9.4	39.3	—	—
419	Qwen2 Instruct 72BAlibaba	11.7	—	—	—	—
420	Ministral 3 3BMistral	11.2	4.8	22.0	204 tok/s	View →
421	Gemini 1.5 Flash-8BGoogle	11.1	—	—	—	—
422	DeepHermes 3 - Mistral 24B Preview (Non-reasoning)Nous Research	10.9	—	—	—	—
423	Jamba 1.7 LargeAI21 Labs	10.9	7.8	2.3	51 tok/s	—
424	Granite 4.0 H SmallIBM	10.8	8.5	13.7	401 tok/s	—
425	Qwen3 Omni 30B A3B InstructAlibaba	10.7	7.2	52.3	108 tok/s	—
426	Jamba 1.5 LargeAI21 Labs	10.7	—	—	—	—
427	DeepSeek-Coder-V2DeepSeek	10.6	—	—	—	—
428	OLMo 2 32BAllen Institute for AI	10.6	2.7	3.3	—	—
429	Hermes 3 - Llama-3.1 70BNous Research	10.6	—	—	32 tok/s	View →
430	Jamba 1.6 LargeAI21 Labs	10.6	—	—	53 tok/s	—
431	Qwen3 8B (Non-reasoning)Alibaba	10.6	7.1	24.3	63 tok/s	—
432	LFM2 24B A2BLiquid AI	10.5	3.6	—	113 tok/s	—
433	Qwen3.5 0.8B (Reasoning)Alibaba	10.5	0.0	—	—	View →
434	Gemini 1.5 Flash (May '24)Google	10.5	—	—	—	—
435	Phi-4Microsoft	10.4	11.2	18.0	39 tok/s	—
436	Nova MicroAmazon	10.3	4.1	6.0	292 tok/s	View →
437	Gemma 3 27B InstructGoogle	10.3	9.6	20.7	—	View →
438	Claude 3 SonnetAnthropic	10.3	—	—	—	—
439	Mistral Small (Sep '24)Mistral	10.2	—	—	151 tok/s	—
440	NVIDIA Nemotron Nano 12B v2 VL (Non-reasoning)NVIDIA	10.1	5.9	26.7	224 tok/s	View →
441	Gemma 3n E4B Instruct Preview (May '25)Google	10.1	—	—	—	—
442	Gemini 1.0 UltraGoogle	10.1	17.6	—	—	—
443	Phi-3 Mini Instruct 3.8BMicrosoft	10.1	3.0	0.3	—	—
444	Phi-4 Multimodal InstructMicrosoft	10.0	—	—	16 tok/s	—
445	Qwen2.5 Coder Instruct 7B Alibaba	10.0	—	—	—	—
446	Qwen3.5 0.8B (Non-reasoning)Alibaba	9.9	1.0	—	89 tok/s	View →
447	Mistral Large (Feb '24)Mistral	9.9	—	—	—	—
448	Mixtral 8x22B InstructMistral	9.8	—	—	—	—
449	Llama 3.2 Instruct 3BMeta	9.7	—	3.3	52 tok/s	—
450	Llama 2 Chat 7BMeta	9.7	—	—	96 tok/s	—
451	Jamba Reasoning 3BAI21 Labs	9.6	2.5	10.7	—	—
452	Qwen3 VL 4B InstructAlibaba	9.6	4.6	37.0	—	—
453	Reka Flash 3Reka AI	9.5	8.9	33.7	95 tok/s	—
454	Qwen1.5 Chat 110BAlibaba	9.5	—	—	—	—
455	Olmo 3 7B ThinkAllen Institute for AI	9.4	7.6	70.7	—	—
456	Claude 2.1Anthropic	9.3	14.0	—	—	—
457	OLMo 2 7BAllen Institute for AI	9.3	1.2	0.7	—	—
458	Molmo 7B-DAllen Institute for AI	9.2	1.2	0.0	—	—
459	Ling-mini-2.0InclusionAI	9.2	5.0	49.3	—	—
460	Claude 2.0Anthropic	9.1	12.9	—	—	—
461	DeepSeek R1 Distill Qwen 1.5BDeepSeek	9.1	—	22.0	—	—
462	DeepSeek-V2-ChatDeepSeek	9.1	—	—	—	—
463	GPT-3.5 TurboOpenAI	9.0	10.7	—	132 tok/s	—
464	Mistral Small (Feb '24)Mistral	9.0	—	—	157 tok/s	—
465	Mistral MediumMistral	9.0	—	—	65 tok/s	—
466	Llama 3 Instruct 70BMeta	8.9	6.8	—	44 tok/s	—
467	Gemma 3 12B InstructGoogle	8.8	6.3	18.3	—	—
468	LFM 40BLiquid AI	8.8	—	—	—	—
469	Arctic InstructSnowflake	8.8	—	—	—	—
470	Qwen Chat 72BAlibaba	8.8	—	—	—	—
471	Llama 3.2 Instruct 11B (Vision)Meta	8.7	4.2	1.7	87 tok/s	—
472	PALM-2Google	8.6	4.6	—	—	—
473	Granite 4.1 3BIBM	8.5	5.5	—	—	—
474	Gemini 1.0 ProGoogle	8.5	—	—	—	—
475	DeepSeek Coder V2 Lite InstructDeepSeek	8.5	—	—	—	—
476	Phi-4 Mini InstructMicrosoft	8.4	3.6	6.7	27 tok/s	—
477	Llama 2 Chat 70BMeta	8.4	—	—	—	—
478	Llama 2 Chat 13BMeta	8.4	—	—	—	—
479	DeepSeek LLM 67B Chat (V1)DeepSeek	8.4	—	—	—	—
480	Sarvam M (Reasoning)Sarvam	8.4	7.5	—	142 tok/s	—
481	Exaone 4.0 1.2B (Reasoning)LG AI Research	8.3	3.1	50.3	—	—
482	OpenChat 3.5 (1210)OpenChat	8.3	—	—	—	—
483	DBRX InstructDatabricks	8.3	—	—	—	—
484	Command-R+ (Apr '24)Cohere	8.3	—	—	—	—
485	LFM2.5-1.2B-ThinkingLiquid AI	8.1	1.4	—	—	—
486	Olmo 3 7B InstructAllen Institute for AI	8.1	3.4	41.3	—	—
487	Exaone 4.0 1.2B (Non-reasoning)LG AI Research	8.1	2.5	24.0	—	—
488	Jamba 1.7 MiniAI21 Labs	8.1	3.1	0.3	—	—
489	LFM2 2.6BLiquid AI	8.0	1.4	8.3	—	—
490	LFM2.5-1.2B-InstructLiquid AI	8.0	0.8	—	—	—
491	Granite 4.0 H 1BIBM	8.0	2.7	6.3	—	—
492	Jamba 1.5 MiniAI21 Labs	8.0	—	—	—	—
493	Qwen3 1.7B (Reasoning)Alibaba	8.0	1.4	38.7	—	—
494	Jamba 1.6 MiniAI21 Labs	7.9	—	—	184 tok/s	—
495	Gemma 3 270MGoogle	7.7	0.0	2.3	—	—
496	Granite 4.0 MicroIBM	7.7	5.0	6.0	—	—
497	Apertus 70B InstructSwiss AI Initiative	7.7	1.9	—	—	—
498	Mixtral 8x7B InstructMistral	7.7	—	—	—	—
499	DeepHermes 3 - Llama-3.1 8B Preview (Non-reasoning)Nous Research	7.6	—	—	—	—
500	Llama 65BMeta	7.4	—	—	—	—
501	Qwen Chat 14BAlibaba	7.4	—	—	—	—
502	Claude InstantAnthropic	7.4	7.8	—	—	—
503	Mistral 7B InstructMistral	7.4	—	—	106 tok/s	—
504	Command-R (Mar '24)Cohere	7.4	—	—	—	—
505	Molmo2-8BAllen Institute for AI	7.3	4.4	—	—	—
506	Granite 4.0 1BIBM	7.3	2.9	6.3	—	—
507	LFM2 8B A1BLiquid AI	7.0	2.3	25.3	—	—
508	Granite 3.3 8B (Non-reasoning)IBM	7.0	3.4	6.7	467 tok/s	—
509	Qwen3 1.7B (Non-reasoning)Alibaba	6.8	2.3	7.3	—	—
510	Qwen3 0.6B (Reasoning)Alibaba	6.5	0.9	18.0	—	—
511	Llama 3 Instruct 8BMeta	6.4	4.0	—	88 tok/s	—
512	Gemma 3n E4B InstructGoogle	6.4	4.2	14.3	52 tok/s	—
513	Llama 3.2 Instruct 1BMeta	6.3	0.6	0.0	93 tok/s	—
514	Gemma 3 4B InstructGoogle	6.3	2.9	12.7	—	—
515	LFM2 1.2BLiquid AI	6.3	0.8	3.3	—	—
516	LFM2.5-VL-1.6BLiquid AI	6.2	1.0	—	—	—
517	Granite 4.0 350MIBM	6.1	0.3	0.0	—	—
518	Apertus 8B InstructSwiss AI Initiative	5.9	1.4	—	—	—
519	Qwen3 0.6B (Non-reasoning)Alibaba	5.7	1.4	10.3	—	—
520	Gemma 3 1B InstructGoogle	5.6	0.2	3.3	—	—
521	Granite 4.0 H 350MIBM	5.4	0.6	1.3	—	—
522	Gemma 3n E2B InstructGoogle	4.8	2.2	10.3	—	—
523	Tiny Aya GlobalCohere	4.7	1.2	—	124 tok/s	—

Benchmarks via Artificial Analysis. Indices are composite scores (0–100). "On Opper" column links to the model's detail page when we host it.