Model rankings

Rank	Change	Model	Intelligence	Price $/M	Speed tok/s	First token
1	—	Kimi K2.6OpenKimi · 1 more variant	53.9	$1.7	42	2.9 s
2	—	MiMo-V2.5-ProOpenXiaomi · 1 more variant	53.8	$1.5	56	3.5 s
3	—	DeepSeek V4 Pro (Reasoning, Max Effort)OpenDeepSeek · 2 more variants	51.5	$2.2	34	1.7 s
4	—	GLM-5.1 (Reasoning)OpenZ AI · 1 more variant	51.4	$2.2	49	1.9 s
5	—	GLM-5 (Reasoning)OpenZ AI · 1 more variant	49.8	$1.6	64	1.6 s
6	—	MiniMax-M2.7OpenMiniMax	49.6	$0.52	48	2.0 s
7	—	MiMo-V2.5OpenXiaomi	49.0	$0.72	93	2.7 s
8	—	Kimi K2.5 (Reasoning)OpenKimi · 1 more variant	46.8	$1.1	49	3.1 s
9	—	DeepSeek V4 Flash (Reasoning, Max Effort)OpenDeepSeek · 2 more variants	46.5	$0.18	72	1.2 s
10	—	Qwen3.6 27B (Reasoning)OpenAlibaba · 1 more variant	45.8	$1.4	61	3.6 s
11	—	Qwen3.5 397B A17B (Reasoning)OpenAlibaba · 1 more variant	45.1	$1.4	53	2.4 s
12	—	Qwen3.6 35B A3B (Reasoning)OpenAlibaba · 1 more variant	43.5	$0.56	194	2.6 s
13	—	GLM-4.7 (Reasoning)OpenZ AI · 1 more variant	42.1	$1.0	99	1.3 s
14	—	Qwen3.5 27B (Reasoning)OpenAlibaba · 1 more variant	42.1	$0.83	88	5.7 s
15	—	MiniMax-M2.5OpenMiniMax	41.9	$0.52	82	1.8 s
16	—	Hy3-preview (Reasoning)OpenTencent · 1 more variant	41.9	$0.000	83	3.3 s
17	—	DeepSeek V3.2 (Reasoning)OpenDeepSeek · 1 more variant	41.7	$0.34	—	—
18	—	Qwen3.5 122B A10B (Reasoning)OpenAlibaba · 1 more variant	41.6	$1.1	151	2.5 s
19	—	MiMo-V2-Flash (Feb 2026)OpenXiaomi · 2 more variants	41.5	$0.15	134	2.3 s
20	—	Kimi K2 ThinkingOpenKimi	40.9	$1.1	120	1.6 s
21	—	MiniMax-M2.1OpenMiniMax	39.4	$0.52	77	2.0 s
22	—	Mistral Medium 3.5OpenMistral	39.2	$3.0	145	1.8 s
23	—	Gemma 4 31B (Reasoning)OpenGoogle · 1 more variant	39.2	$0.000	36	1.7 s
24	—	Step 3.5 FlashOpenStepFun	37.8	$0.15	182	1.0 s
25	—	Qwen3.5 35B A3B (Reasoning)OpenAlibaba · 1 more variant	37.1	$0.69	167	2.4 s
26	—	MiniMax-M2OpenMiniMax	36.1	$0.52	70	1.8 s
27	—	NVIDIA Nemotron 3 Super 120B A12B (Reasoning)OpenNVIDIA	36.0	$0.41	174	1.2 s
28	—	DeepSeek V3.1 Terminus (Reasoning)OpenDeepSeek · 1 more variant	33.9	$1.9	—	—
29	—	Ling-2.6-1TOpenInclusionAI	33.6	$0.85	—	—
30	—	gpt-oss-120B (high)OpenOpenAI · 1 more variant	33.3	$0.26	243	0.9 s
31	—	DeepSeek V3.2 Exp (Reasoning)OpenDeepSeek · 1 more variant	32.9	$0.31	—	—
32	—	GLM-4.6 (Reasoning)OpenZ AI · 1 more variant	32.5	$0.96	38	1.4 s
33	—	Qwen3.5 9B (Reasoning)OpenAlibaba · 1 more variant	32.4	$0.11	58	0.7 s
34	—	K-EXAONE (Reasoning)OpenLG AI Research · 1 more variant	32.1	—	—	—
35	—	Trinity Large ThinkingOpenArcee AI	31.9	$0.40	118	0.9 s
36	—	Gemma 4 26B A4B (Reasoning)OpenGoogle · 1 more variant	31.2	$0.20	—	—
37	—	Kimi K2 0905OpenKimi · 1 more variant	30.9	$1.1	18	2.9 s
38	—	EXAONE 4.5 33BOpenLG AI Research	30.2	—	—	—
39	—	GLM-4.7-Flash (Reasoning)OpenZ AI · 1 more variant	30.2	$0.15	80	1.1 s
40	—	Qwen3 235B A22B 2507 (Reasoning)OpenAlibaba · 2 more variants	29.5	$0.84	48	2.9 s
41	—	DeepSeek V3.2 SpecialeOpenDeepSeek	29.4	—	—	—
42	—	Nemotron Cascade 2 30B A3BOpenNVIDIA	28.4	—	—	—
43	—	Qwen3 Coder NextOpenAlibaba	28.3	$0.56	144	1.1 s
44	—	DeepSeek V3.1 (Non-reasoning)OpenDeepSeek · 1 more variant	28.1	$0.83	—	—
45	—	Mistral Small 4 (Reasoning)OpenMistral · 1 more variant	27.8	$0.26	150	0.7 s
46	—	Qwen3 VL 235B A22B (Reasoning)OpenAlibaba	27.6	$2.2	31	3.5 s
47	—	Qwen3.5 4B (Reasoning)OpenAlibaba · 1 more variant	27.1	$0.060	203	0.5 s
48	—	DeepSeek R1 0528 (May '25)OpenDeepSeek · 1 more variant	27.1	$2.1	—	—
49	—	Qwen3 Next 80B A3B (Reasoning)OpenAlibaba	26.7	$1.9	157	2.2 s
50	—	GLM-4.5 (Reasoning)OpenZ AI	26.4	$1.0	43	3.1 s
51	—	Ling 2.6 FlashOpenInclusionAI	26.2	$0.15	206	1.6 s
52	—	Seed-OSS-36B-InstructOpenByteDance Seed	25.2	$0.30	37	3.2 s
53	—	Qwen3 235B A22B 2507 InstructOpenAlibaba	25.0	$0.36	56	2.4 s
54	—	Qwen3 Coder 480B A35B InstructOpenAlibaba	24.8	$0.68	55	3.1 s
55	—	Qwen3 VL 32B (Reasoning)OpenAlibaba	24.7	$2.6	87	2.7 s
56	—	gpt-oss-20B (high)OpenOpenAI · 1 more variant	24.5	$0.088	262	0.7 s
57	—	MiniMax M1 80kOpenMiniMax	24.4	$0.96	—	—
58	—	NVIDIA Nemotron 3 Nano 30B A3B (Reasoning)OpenNVIDIA · 1 more variant	24.3	$0.096	188	1.5 s
59	—	LongCat Flash LiteOpenLongCat	23.9	$0.000	139	8.9 s
60	—	GLM-4.6V (Reasoning)OpenZ AI · 1 more variant	23.4	$0.45	39	2.8 s
61	—	GLM-4.5-AirOpenZ AI	23.2	$0.37	77	2.9 s
62	—	Mistral Large 3OpenMistral	22.8	$0.75	48	1.1 s
63	—	Ring-1TOpenInclusionAI	22.8	—	—	—
64	—	Qwen3 30B A3B 2507 (Reasoning)OpenAlibaba · 2 more variants	22.4	$0.67	140	2.5 s
65	—	DeepSeek V3 0324OpenDeepSeek · 1 more variant	22.3	$1.2	—	—
66	—	INTELLECT-3OpenPrime Intellect	22.2	—	—	—
67	—	Devstral 2OpenMistral	22.0	$0.000	61	1.2 s
68	—	Solar Open 100B (Reasoning)OpenUpstage	21.7	—	—	—
69	New	Nemotron 3 Nano Omni 30B A3B ReasoningOpenNVIDIA	21.4	$0.13	322	0.9 s
70	▼ 1	MiniMax M1 40kOpenMiniMax	20.9	—	—	—
71	▼ 1	Qwen3 VL 235B A22B InstructOpenAlibaba	20.8	$0.70	47	2.9 s
72	▼ 1	Qwen3 Next 80B A3B InstructOpenAlibaba	20.1	$0.88	148	2.3 s
73	▼ 1	Qwen3 Coder 30B A3B InstructOpenAlibaba	20.0	$0.35	106	2.7 s
74	▼ 1	QwQ 32BOpenAlibaba	19.7	$0.75	29	2.1 s
75	▼ 1	Qwen3 VL 30B A3B (Reasoning)OpenAlibaba	19.7	$0.34	110	2.2 s
76	▼ 1	Devstral Small 2OpenMistral	19.5	$0.000	61	1.1 s
77	▼ 1	Ling-1TOpenInclusionAI	19.0	—	—	—
78	▼ 1	Gemma 4 E4B (Reasoning)OpenGoogle · 1 more variant	18.8	$0.54	56	1.7 s
79	▼ 1	Llama Nemotron Super 49B v1.5 (Reasoning)OpenNVIDIA · 1 more variant	18.7	$0.18	46	1.6 s
80	▼ 1	Llama 3.3 Nemotron Super 49B v1 (Reasoning)OpenNVIDIA · 1 more variant	18.5	—	—	—
81	▼ 1	Llama 4 MaverickOpenMeta	18.4	$0.48	107	1.0 s
82	▼ 1	Qwen3 4B 2507 (Reasoning)OpenAlibaba · 2 more variants	18.2	—	—	—
83	▼ 1	Magistral Small 1.2OpenMistral	18.2	$0.75	106	0.8 s
83	▼ 1	Sarvam 105B (high)OpenSarvam	18.2	$0.000	106	2.3 s
85	▼ 1	Devstral Small (May '25)OpenMistral · 1 more variant	18.0	—	—	—
86	▼ 1	Llama 3.1 Instruct 405BOpenMeta	17.4	$3.7	30	2.4 s
87	▼ 1	Qwen3 VL 32B InstructOpenAlibaba	17.2	$1.2	74	2.8 s
88	▼ 1	DeepSeek R1 Distill Qwen 32BOpenDeepSeek	17.2	—	—	—
89	▼ 1	Magistral Small 1OpenMistral	16.8	—	—	—
90	▼ 1	EXAONE 4.0 32B (Reasoning)OpenLG AI Research · 1 more variant	16.7	—	—	—
91	▼ 1	Qwen3 VL 8B (Reasoning)OpenAlibaba	16.7	$0.66	110	2.4 s
92	▼ 1	Qwen3 32B (Reasoning)OpenAlibaba · 1 more variant	16.5	$0.28	82	2.6 s
93	▼ 1	DeepSeek R1 0528 Qwen3 8BOpenDeepSeek	16.4	—	—	—
94	▼ 1	Qwen3.5 2B (Reasoning)OpenAlibaba · 1 more variant	16.3	$0.040	—	—
95	▼ 1	Qwen3 14B (Reasoning)OpenAlibaba · 1 more variant	16.2	$0.73	59	2.9 s
96	▼ 1	Nanbeige4.1-3BOpenNanbeige	16.1	—	—	—
97	▼ 1	Qwen3 VL 30B A3B InstructOpenAlibaba	16.1	$0.30	108	2.3 s
98	▼ 1	Ministral 3 14BOpenMistral	16.0	$0.20	136	0.7 s
99	▼ 1	DeepSeek R1 Distill Llama 70BOpenDeepSeek	16.0	$0.79	42	1.8 s
100	▼ 1	DeepSeek R1 Distill Qwen 14BOpenDeepSeek	15.8	—	—	—
101	▼ 1	Falcon-H1R-7BOpenTII UAE	15.8	—	—	—
102	▼ 1	Ling-flash-2.0OpenInclusionAI	15.7	$0.25	82	2.7 s
103	▼ 1	Qwen3 Omni 30B A3B (Reasoning)OpenAlibaba	15.6	$0.43	79	1.9 s
104	▼ 1	Qwen2.5 Instruct 72BOpenAlibaba	15.6	$0.37	28	3.6 s
105	▼ 1	Step3 VL 10BOpenStepFun	15.5	—	—	—
106	▼ 1	Gemma 4 E2B (Reasoning)OpenGoogle · 1 more variant	15.2	—	—	—
107	▼ 1	QwQ 32B-PreviewOpenAlibaba	15.2	—	—	—
108	▼ 1	GLM-4.5V (Reasoning)OpenZ AI · 1 more variant	15.1	$0.90	41	1.7 s
108	▼ 1	Mistral Large 2 (Nov '24)OpenMistral · 1 more variant	15.1	$3.0	30	1.8 s
110	▼ 1	Mistral Small 3.2OpenMistral	15.1	$0.13	138	0.6 s
111	▼ 1	Llama 3.1 Nemotron Ultra 253B v1 (Reasoning)OpenNVIDIA	15.0	$0.90	41	2.6 s
112	▼ 1	Qwen3 30B A3B 2507 InstructOpenAlibaba	15.0	$0.21	116	2.0 s
113	▼ 1	ERNIE 4.5 300B A47BOpenBaidu	15.0	$0.49	22	3.6 s
114	▼ 1	NVIDIA Nemotron Nano 12B v2 VL (Reasoning)OpenNVIDIA · 1 more variant	14.9	$0.30	128	1.1 s
115	▼ 1	Ministral 3 8BOpenMistral	14.8	$0.15	152	0.6 s
116	▼ 1	NVIDIA Nemotron Nano 9B V2 (Reasoning)OpenNVIDIA · 1 more variant	14.8	$0.070	121	1.1 s
117	▼ 1	Granite 4.1 30BOpenIBM	14.7	—	—	—
118	▼ 1	NVIDIA Nemotron 3 Nano 4BOpenNVIDIA	14.7	—	—	—
119	▼ 1	Llama 3.3 Instruct 70BOpenMeta	14.5	$0.62	77	1.5 s
120	▼ 1	Mistral Small 3.1OpenMistral	14.5	$0.14	135	0.7 s
121	▼ 1	Llama 3.1 Nemotron Nano 4B v1.1 (Reasoning)OpenNVIDIA	14.4	—	—	—
122	▼ 1	Kimi Linear 48B A3B InstructOpenKimi	14.4	—	—	—
123	▼ 1	Qwen3 VL 8B InstructOpenAlibaba	14.3	$0.31	119	2.4 s
124	▼ 1	Ring-flash-2.0OpenInclusionAI	14.0	$0.25	88	2.4 s
125	▼ 1	Pixtral LargeOpenMistral	14.0	$3.0	54	1.7 s
126	▼ 1	Grok 2 (Dec '24)OpenxAI	13.9	—	—	—
127	▼ 1	Qwen3 VL 4B (Reasoning)OpenAlibaba	13.7	—	—	—
128	▼ 1	Llama 4 ScoutOpenMeta	13.5	$0.29	142	0.8 s
129	▼ 1	Command AOpenCohere	13.5	$4.4	38	2.4 s
130	▼ 1	Llama 3.1 Nemotron Instruct 70BOpenNVIDIA	13.4	$1.2	37	1.7 s
131	▼ 1	Qwen2.5 Instruct 32BOpenAlibaba	13.2	—	—	—
132	▼ 1	Qwen3 8B (Reasoning)OpenAlibaba · 1 more variant	13.2	$0.37	81	2.3 s
133	▼ 1	Qwen3 4B 2507 InstructOpenAlibaba	12.9	—	—	—
134	▼ 1	Qwen2.5 Coder Instruct 32BOpenAlibaba	12.9	—	—	—
135	▼ 1	Mistral Small 3OpenMistral	12.7	$0.10	133	0.7 s
136	▼ 1	DeepSeek-V2.5 (Dec '24)OpenDeepSeek · 1 more variant	12.5	—	—	—
137	▼ 1	Llama 3.1 Instruct 70BOpenMeta	12.5	$0.56	33	1.7 s
138	▼ 1	Granite 4.1 8BOpenIBM	12.4	$0.063	92	0.9 s
139	▼ 1	Sarvam 30B (high)OpenSarvam	12.3	$0.000	260	1.9 s
140	▼ 1	DeepSeek R1 Distill Llama 8BOpenDeepSeek	12.1	—	—	—
141	▼ 1	R1 1776OpenPerplexity	12.0	—	—	—
142	▼ 1	Llama 3.2 Instruct 90B (Vision)OpenMeta	11.9	$1.4	58	1.2 s
142	▼ 1	Solar MiniOpenUpstage	11.9	$0.15	—	—
144	▼ 1	Llama 3.1 Instruct 8BOpenMeta	11.8	$0.10	155	0.9 s
145	▼ 1	Grok-1OpenxAI	11.7	—	—	—
146	▼ 1	Qwen2 Instruct 72BOpenAlibaba	11.7	—	—	—
147	▼ 1	Ministral 3 3BOpenMistral	11.2	$0.10	278	0.5 s
148	▼ 1	Jamba 1.7 LargeOpenAI21 Labs	10.9	$3.5	62	1.7 s
149	▼ 1	Granite 4.0 H SmallOpenIBM	10.8	$0.11	243	10 s
150	▼ 1	Jamba 1.5 LargeOpenAI21 Labs	10.7	$3.5	—	—
151	▼ 1	Qwen3 Omni 30B A3B InstructOpenAlibaba	10.7	$0.43	95	2.0 s
152	▼ 1	DeepSeek-Coder-V2OpenDeepSeek	10.6	—	—	—
153	▼ 1	Jamba 1.6 LargeOpenAI21 Labs	10.6	$3.5	60	1.5 s
154	▼ 1	Qwen3.5 0.8B (Reasoning)OpenAlibaba · 1 more variant	10.5	$0.020	—	—
155	▼ 1	LFM2 24B A2BOpenLiquid AI	10.5	$0.053	73	0.6 s
156	▼ 1	Phi-4OpenMicrosoft	10.4	$0.22	35	2.1 s
157	▼ 1	Gemma 3 27B InstructOpenGoogle	10.3	$0.14	26	2.0 s
158	▼ 1	Mistral Small (Sep '24)OpenMistral	10.2	$0.30	130	0.8 s
159	▼ 1	Phi-3 Mini Instruct 3.8BOpenMicrosoft	10.1	—	—	—
160	▼ 1	Gemma 3n E4B Instruct Preview (May '25)OpenGoogle	10.1	—	—	—
161	▼ 1	Phi-4 Multimodal InstructOpenMicrosoft	10.0	$0.000	17	1.1 s
162	▼ 1	Qwen2.5 Coder Instruct 7BOpenAlibaba	10.0	—	—	—
163	▼ 1	Mixtral 8x22B InstructOpenMistral	9.8	—	—	—
164	▼ 1	Llama 2 Chat 7BOpenMeta	9.7	$0.10	—	—
165	▼ 1	Llama 3.2 Instruct 3BOpenMeta	9.7	$0.15	51	1.1 s
166	▼ 1	Jamba Reasoning 3BOpenAI21 Labs	9.6	—	—	—
167	▼ 1	Qwen3 VL 4B InstructOpenAlibaba	9.6	—	—	—
168	▼ 1	Qwen1.5 Chat 110BOpenAlibaba	9.5	—	—	—
169	▼ 1	Reka Flash 3OpenReka AI	9.5	$0.35	—	—
170	▼ 1	Ling-mini-2.0OpenInclusionAI	9.2	—	—	—
171	▼ 1	DeepSeek R1 Distill Qwen 1.5BOpenDeepSeek	9.1	—	—	—
172	▼ 1	DeepSeek-V2-ChatOpenDeepSeek	9.1	—	—	—
173	▼ 1	Llama 3 Instruct 70BOpenMeta	8.9	$1.2	—	—
174	▼ 1	Arctic InstructOpenSnowflake	8.8	—	—	—
174	▼ 1	Qwen Chat 72BOpenAlibaba	8.8	—	—	—
176	▼ 1	Gemma 3 12B InstructOpenGoogle	8.8	$0.14	25	2.4 s
177	▼ 1	Llama 3.2 Instruct 11B (Vision)OpenMeta	8.7	$0.25	49	0.9 s
178	▼ 1	Granite 4.1 3BOpenIBM	8.5	—	—	—
179	▼ 1	DeepSeek Coder V2 Lite InstructOpenDeepSeek	8.5	—	—	—
180	▼ 1	Phi-4 Mini InstructOpenMicrosoft	8.4	$0.000	44	0.8 s
180	▼ 1	Sarvam M (Reasoning)OpenSarvam	8.4	$0.000	—	—
182	▼ 1	DeepSeek LLM 67B Chat (V1)OpenDeepSeek	8.4	—	—	—
182	▼ 1	Llama 2 Chat 70BOpenMeta	8.4	—	—	—
184	▼ 1	Llama 2 Chat 13BOpenMeta	8.4	—	—	—
185	▼ 1	Command-R+ (Apr '24)OpenCohere · 1 more variant	8.3	$6.0	—	—
186	▼ 1	OpenChat 3.5 (1210)OpenOpenChat	8.3	—	—	—
187	▼ 1	DBRX InstructOpenDatabricks	8.3	—	—	—
188	▼ 1	Exaone 4.0 1.2B (Reasoning)OpenLG AI Research · 1 more variant	8.3	—	—	—
189	▼ 1	LFM2.5-1.2B-ThinkingOpenLiquid AI	8.1	—	—	—
190	▼ 1	Jamba 1.7 MiniOpenAI21 Labs	8.1	—	—	—
191	▼ 1	LFM2 2.6BOpenLiquid AI	8.0	$0.000	—	—
191	▼ 1	LFM2.5-1.2B-InstructOpenLiquid AI	8.0	$0.000	—	—
193	▼ 1	Jamba 1.5 MiniOpenAI21 Labs	8.0	$0.25	—	—
194	▼ 1	Granite 4.0 H 1BOpenIBM	8.0	—	—	—
195	▼ 1	Qwen3 1.7B (Reasoning)OpenAlibaba · 1 more variant	8.0	$0.40	127	1.9 s
196	▼ 1	Jamba 1.6 MiniOpenAI21 Labs	7.9	$0.25	183	0.9 s
197	▼ 1	Mixtral 8x7B InstructOpenMistral	7.7	$0.51	—	—
198	▼ 1	Gemma 3 270MOpenGoogle	7.7	—	—	—
199	▼ 1	Apertus 70B InstructOpenSwiss AI Initiative	7.7	$1.3	—	—
200	▼ 1	Granite 4.0 MicroOpenIBM	7.7	—	—	—
201	▼ 1	Llama 65BOpenMeta	7.4	—	—	—
201	▼ 1	Mistral 7B InstructOpenMistral	7.4	$0.21	143	0.6 s
201	▼ 1	Qwen Chat 14BOpenAlibaba	7.4	—	—	—
204	▼ 1	Granite 4.0 1BOpenIBM	7.3	—	—	—
205	▼ 1	LFM2 8B A1BOpenLiquid AI	7.0	$0.000	—	—
206	▼ 1	Granite 3.3 8B (Non-reasoning)OpenIBM	7.0	$0.085	290	27 s
207	▼ 1	Qwen3 0.6B (Reasoning)OpenAlibaba · 1 more variant	6.5	$0.40	188	1.7 s
208	▼ 1	Gemma 3n E4B InstructOpenGoogle	6.4	$0.025	24	2.2 s
208	▼ 1	Llama 3 Instruct 8BOpenMeta	6.4	$0.070	—	—
210	▼ 1	LFM2 1.2BOpenLiquid AI	6.3	$0.000	—	—
211	▼ 1	Gemma 3 4B InstructOpenGoogle	6.3	$0.050	25	1.8 s
212	▼ 1	Llama 3.2 Instruct 1BOpenMeta	6.3	$0.050	95	0.9 s
213	▼ 1	LFM2.5-VL-1.6BOpenLiquid AI	6.2	$0.000	—	—
214	▼ 1	Granite 4.0 350MOpenIBM	6.1	—	—	—
215	▼ 1	Apertus 8B InstructOpenSwiss AI Initiative	5.9	$0.13	—	—
216	▼ 1	Gemma 3 1B InstructOpenGoogle	5.6	$0.000	60	0.8 s
217	▼ 1	Granite 4.0 H 350MOpenIBM	5.4	—	—	—
218	▼ 1	Gemma 3n E2B InstructOpenGoogle	4.8	$0.000	—	—
219	▼ 1	Tiny Aya GlobalOpenCohere	4.7	$0.000	—	—

169 more models

Intelligence over time

Each maker's best model at the time on the Artificial Analysis Intelligence Index, by the month it was released.

Gained most · Anthropic +57.2Highest now · Anthropic 64.6Gap between highest and lowest 18.1 ptsTime range Nov 2025 – Nov 2026

64.6Anthropic60.0Google56.9OpenAI50.1Meta46.5Kimi