Ran it, or it didn’t happen.

Benchmark open models on your own hardware. Runs count once their event log checks out.

Recent verified runsWhereSpeed

PassedFailedNot supported

Cloud leaderboard

Every model runs on the same pinned machine. 17 models, 70 runs.

Suite: [email protected]
Machine: NVIDIA L40S · 48 GB
Engine: ollama 0.40.14

#	Model	Pass rate				Speed	Runs
1	Qwen3.5-35B-A3BQ4_K_M	96%	9/9	8/9	9/9	201.4 tok/s	3
2	Qwen3.5-27BQ4_K_M	93%	8/9	9/9	8/9	35.5 tok/s	3
3	Qwen3-30B-A3B-Instruct-2507Q4_K_M	92%	11/12	10/12	12/12	187.6 tok/s	4
4	GLM-4.7-FlashQ4_K_M	89%	9/9	6/9	9/9	201.7 tok/s	3
5	Qwen3-Coder-30B-A3B-InstructQ4_K_M	89%	9/9	7/9	8/9	186.4 tok/s	3
6	Qwen3-8BQ4_K_M	89%	9/12	11/12	12/12	120.7 tok/s	4
7	Qwen3-14BQ4_K_M	87%	13/15	11/15	15/15	67.1 tok/s	5
8	gemma-3-27b-itQ4_K_M	86%	17/18	Not supported	14/18	36.2 tok/s	6
9	gpt-oss-20bMXFP4	85%	9/9	9/9	5/9	173.8 tok/s	3
10	gemma-3-4b-itQ4_K_M	83%	11/15	Not supported	14/15	232.7 tok/s	5
11	Qwen3.5-2BQ4_K_M	81%	9/12	9/12	11/12	298.1 tok/s	4
12	Qwen3-4BQ4_K_M	78%	6/9	6/9	9/9	247.9 tok/s	3
13	gemma-3-12b-itQ4_K_M	75%	8/12	Not supported	10/12	80.9 tok/s	4
14	Llama-3.1-8B-InstructQ4_K_M	71%	8/15	9/15	15/15	124.4 tok/s	5
15	Phi-3.5-mini-instructQ4_K_M	64%	9/18	Not supported	14/18	261.3 tok/s	6
16	MiniCPM5-2BQ4_K_M	52%	4/9	3/9	7/9	301.4 tok/s	3
17	Mistral-7B-Instruct-v0.2Q4_K_M	47%	6/18	Not supported	11/18	137.1 tok/s	6

Ranked by checks passed. A task a model can’t do is shown, not counted as a zero.

Open model leaderboard

144 open-weight models in instructions ranked by LMArena score · updated Nov 6, 2026

Top rated

Most downloaded

Newest

All sizes ≤3B 3–9B 9–20B 20–40B 40–80B >80B

#	Model	Score	Votes	Params	License	Downloads	Likes
1	Kimi-K3moonshotai	1484	13.6K	2.8T	Custom	771.6K	11,950
2	GLM-5.3zai-org	1483	10.3K	753B	Custom	2.1M	2,395
3	GLM-5.3-Flashzai-org	1480	17.4K	321B	MIT	6.4M	3,088
4	MiMo-V2.5-ProXiaomiMiMo	1475	26.1K	1.0T	MIT	9.9K	771
5	GLM-5.2zai-org	1466	19.1K	753B	MIT	318.5K	5,205
6	GLM-5.1zai-org	1452	23.9K	754B	MIT	892.7K	1,850
7	Kimi-K2.6moonshotai	1451	13.2K	1.0T	Modified MIT	237.2K	1,641
8	DeepSeek-V4-Prodeepseek-ai	1449	3.5K	1.6T	MIT	335.6K	5,699
9	MiMo-V2.5XiaomiMiMo	1435	16K	311B	MIT	110.3K	434
10	gemma-4-31Bgoogle	1432	1.8K	31.3B	Apache 2.0	57K	589
11	Qwen3.8-27BQwen	1432	11.5K	27.8B	Apache 2.0	7.8M	19,348
12	Kimi-K2.5moonshotai	1431	24.2K	1.0T	Modified MIT	395K	2,901
13	Beam-1.0-501B-A23Breflection	1430	1.6K	501B	Apache 2.0	11.8K	311
14	Inklingthinkingmachines	1429	18.1K	952B	Apache 2.0	1.8M	1,842
15	GLM-5zai-org	1428	9.6K	754B	MIT	1.4M	2,141
16	MiniMax-M3MiniMaxAI	1428	25.5K	427B	Custom	155.9K	1,600
17	Hy3tencent	1426	5.4K	299B	Apache 2.0	1.2M	990
18	Qwen3.5-397B-A17BQwen	1424	34.1K	403B	Apache 2.0	857.2K	1,581
19	DeepSeek-V4-Flashdeepseek-ai	1421	17.9K	291B	MIT	1.1M	2,378
20	gemma-4-26B-A4Bgoogle	1420	1.7K	25.8B	Apache 2.0	74.1K	465
21	NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16nvidia	1417	3.9K	561B	OpenMDW 1.1	1.1M	384
22	Mistral-Medium-3.5-128Bmistralai	1415	3.8K	128B	Custom	17.2K	498
23	DeepSeek-V3.2deepseek-ai	1413	13.6K	685B	MIT	7.5M	1,596
24	GLM-4.7zai-org	1412	3.3K	358B	MIT	66.6K	2,077
25	GLM-4.6zai-org	1410	10.3K	357B	MIT	28.1K	1,247
26	Qwen3-235B-A22B-Instruct-2507Qwen	1408	27.8K	235B	Apache 2.0	84.8K	814
27	Qwen3-VL-235B-A22B-InstructQwen	1406	2.9K	236B	Apache 2.0	327K	431
28	MiniMax-M2.7MiniMaxAI	1405	30.5K	229B	Custom	307.1K	1,261
29	DeepSeek-V3.1-Terminusdeepseek-ai	1405	858	685B	MIT	9.5K	386
30	Mistral-Large-3-675B-Instruct-2512mistralai	1404	29.3K	—	Apache 2.0	2.5K	261
31	GLM-4.5zai-org	1404	6.2K	358B	MIT	110.4K	1,439
32	Inkling-Smallthinkingmachines	1402	14.2K	266B	Apache 2.0	1.2M	449
33	DeepSeek-V3.2-Expdeepseek-ai	1401	3.4K	685B	MIT	43.6K	1,027
34	DeepSeek-V3.1deepseek-ai	1400	2.8K	685B	MIT	262.8K	864
35	Qwen3.5-122B-A10BQwen	1400	9.6K	125B	Apache 2.0	924.6K	653
36	LongCat-Flash-Chatmeituan-longcat	1398	2.9K	562B	MIT	13.3K	540
37	NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16nvidia	1396	1.5K	124B	OpenMDW 1.1	70.3K	116
38	Qwen3.5-27BQwen	1393	9.2K	27.8B	Apache 2.0	1.7M	1,090
39	MiMo-V2-FlashXiaomiMiMo	1393	14.7K	310B	MIT	22.2K	756
40	Qwen3-Next-80B-A3B-InstructQwen	1388	6.4K	81.3B	Apache 2.0	129.9K	1,072
41	Qwen3-235B-A22B-Thinking-2507Qwen	1386	2.1K	235B	Apache 2.0	8.9K	419
42	Step-3.5-Flashstepfun-ai	1385	19.3K	199B	Apache 2.0	219K	847
43	Hy3-previewtencent	1382	2.3K	299B	Custom	68.3K	314
44	DeepSeek-R1-0528deepseek-ai	1382	4K	685B	MIT	122.3K	2,495
45	Qwen3.5-35B-A3BQwen	1379	9.9K	36.0B	Apache 2.0	1.6M	1,555
46	Muse-Glimmer-30Bmeta-models	1375	1.4K	29.8B	Apache 2.0	187.5K	2,127
47	Qwen3-VL-235B-A22B-ThinkingQwen	1374	2.2K	236B	Apache 2.0	81K	420
48	Qwen3-30B-A3B-Instruct-2507Qwen	1363	6K	30.5B	Apache 2.0	932.4K	855
49	Qwen3-235B-A22BQwen	1362	9.2K	235B	Apache 2.0	308.9K	1,128
50	DeepSeek-R1deepseek-ai	1357	6.4K	684B	MIT	1.6M	14,439
51	Qwen3-Coder-480B-A35B-InstructQwen	1355	6.2K	480B	Apache 2.0	58.6K	1,387
52	GLM-4.5-Airzai-org	1354	8.2K	110B	MIT	1.3M	649
53	MiniMax-M2.5MiniMaxAI	1353	13.7K	229B	Modified MIT	1.2M	1,529
54	GLM-4.6Vzai-org	1351	757	108B	MIT	5.5K	398
55	NVIDIA-Nemotron-3-Super-120B-A12B-BF16nvidia	1347	2.2K	124B	Custom	1.1M	438
56	granite-4.2-30bibm-granite	1346	1.2K	29.3B	Apache 2.0	32.9K	140
57	DeepSeek-V3-0324deepseek-ai	1345	12.4K	685B	MIT	153.8K	3,223
58	Qwen3-Next-80B-A3B-ThinkingQwen	1343	3.5K	81.3B	Apache 2.0	44.4K	503
59	Trinity-Large-Previewarcee-ai	1334	9.7K	399B	OpenMDW 1.1	1.7K	181
60	MiniMax-M2MiniMaxAI	1329	2K	229B	Modified MIT	109.4K	1,514
61	GLM-4.7-Flashzai-org	1327	3.3K	31.2B	MIT	1M	1,896
62	Trinity-Large-Thinkingarcee-ai	1326	10.3K	399B	OpenMDW 1.1	3.2K	190
63	gemma-3-27b-itgoogle	1321	12.5K	27.4B	Gemma	364.5K	2,066
64	NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16nvidia	1320	8.9K	31.6B	OpenMDW 1.1	1.1M	271
65	Ling-flash-2.0inclusionAI	1319	1.8K	103B	MIT	4.4K	222
66	gpt-oss-120bopenai	1318	7.9K	117B	Apache 2.0	2.4M	5,498
67	INTELLECT-3PrimeIntellect	1315	1.5K	107B	MIT	774	218
68	DeepSeek-V3deepseek-ai	1315	8.6K	684B	—	1.6M	4,462
69	Ring-flash-2.0inclusionAI	1313	1.9K	103B	MIT	1.4K	104
70	GLM-4.5Vzai-org	1311	1.3K	108B	MIT	16.2K	726
71	Mistral-Small-3.2-24B-Instruct-2506mistralai	1309	4.4K	24.0B	Apache 2.0	109.2K	644
72	Qwen3-32BQwen	1305	858	32.8B	Apache 2.0	3M	791
73	NVIDIA-Nemotron-3-Nano-30B-A3B-BF16nvidia	1304	4.4K	31.6B	Custom	1.6M	840
74	Olmo-3.1-32B-Instructallenai	1300	3.2K	32.2B	Apache 2.0	20.8K	88
75	granite-4.2-8bibm-granite	1299	1.1K	8.8B	Apache 2.0	69.2K	116
76	gemma-3-12b-itgoogle	1299	1.1K	12.2B	Gemma	365.8K	908
77	QwQ-32BQwen	1297	7.2K	32.8B	Apache 2.0	32K	2,971
78	Qwen3-30B-A3BQwen	1283	6.1K	30.5B	Apache 2.0	1.1M	982
79	DeepSeek-V2.5-1210deepseek-ai	1279	3K	236B	Custom	2.1K	275
80	Olmo-3-32B-Thinkallenai	1274	1.5K	32.2B	Apache 2.0	12.2K	177
81	granite-4.2-3bibm-granite	1273	1.1K	3.7B	Apache 2.0	38.9K	131
82	Molmo2-8Ballenai	1270	220	8.7B	Apache 2.0	44.4K	198
83	granite-4.1-8bibm-granite	1267	1.4K	8.8B	Apache 2.0	98.1K	257
84	Llama-4-Maverick-17B-128E-Instructmeta-llama	1267	10.5K	402B	Llama 4	11.7K	520
85	Mistral-Small-3.1-24B-Instruct-2503mistralai	1263	8K	24.0B	Apache 2.0	168.4K	1,408
86	Llama-3.1-405B-Instructmeta-llama	1258	16.2K	406B	Llama 3.1	9.2K	607
87	gemma-3n-E4B-itgoogle	1255	5K	7.8B	Gemma	12.7K	938
88	Qwen2.5-72B-InstructQwen	1254	16.4K	72.7B	Custom	192.8K	1,043
89	Llama-3.1-Nemotron-70B-Instructnvidia	1252	3K	—	Llama 3.1	74	570
90	DeepSeek-V2.5deepseek-ai	1249	10.2K	236B	Custom	4.5K	750
91	Llama-4-Scout-17B-16E-Instructmeta-llama	1248	7.5K	109B	Llama 4	110.7K	1,387
92	Olmo-3.1-32B-Thinkallenai	1246	2.2K	32.2B	Apache 2.0	7K	120
93	Llama-3.3-70B-Instructmeta-llama	1242	18.8K	70.6B	Llama 3.3	357.4K	3,197
94	gemma-3-4b-itgoogle	1239	1.2K	4.3B	Gemma	1M	1,633
95	gpt-oss-20bopenai	1236	2.5K	20.9B	Apache 2.0	5.9M	5,286
96	Llama-3.1-Tulu-3-70Ballenai	1232	1.2K	70.6B	Llama 3.1	965	61
97	Llama-3.1-70B-Instructmeta-llama	1231	21.9K	70.6B	Llama 3.1	92K	1,027
98	Qwen2.5-Coder-32B-InstructQwen	1223	2.2K	32.8B	Apache 2.0	548.8K	2,206
99	granite-4.0-h-smallibm-granite	1222	1.6K	32.2B	Apache 2.0	23.5K	313
100	Mistral-Small-24B-Instruct-2501mistralai	1214	5.5K	23.6B	Apache 2.0	94.4K	991
101	gemma-2-27b-itgoogle	1206	29.5K	27.2B	Gemma	21.7K	577
102	Nemotron-4-340B-Instructnvidia	1203	7.4K	—	Custom	638	703
103	phi-4microsoft	1201	9.2K	14.7B	MIT	293.7K	2,340
104	aya-expanse-32bCohereLabs	1196	11.3K	32.3B	CC BY-NC 4.0	3K	303
105	Meta-Llama-3-70B-Instructmeta-llama	1194	56.6K	70.6B	Llama 3	37.8K	1,537
106	Qwen2-72B-InstructQwen	1181	14.2K	72.7B	Tongyi Qianwen	24.3K	724
107	DeepSeek-Coder-V2-Instructdeepseek-ai	1180	5.6K	236B	Custom	3.8K	718
108	gemma-2-9b-itgoogle	1178	21.4K	9.2B	Gemma	113.6K	1,166
109	Llama-3.1-Tulu-3-8Ballenai	1174	1.2K	8.0B	Llama 3.1	5K	178
110	Llama-3.1-8B-Instructmeta-llama	1159	19.8K	8.0B	Llama 3.1	4.3M	8,836
111	Qwen1.5-110B-ChatQwen	1158	9.5K	111B	Tongyi Qianwen	2.7K	139
112	QwQ-32B-PreviewQwen	1156	1.3K	32.8B	Apache 2.0	7.1K	1,747
113	aya-expanse-8bCohereLabs	1155	4K	8.0B	CC BY-NC 4.0	22.9K	463
114	Mixtral-8x22B-Instruct-v0.1mistralai	1147	18.5K	141B	Apache 2.0	38.7K	780
115	Qwen1.5-72B-ChatQwen	1141	13.8K	72.3B	Tongyi Qianwen	11.2K	224
116	granite-3.1-8b-instructibm-granite	1131	1.3K	8.2B	Apache 2.0	27.4K	169
117	Meta-Llama-3-8B-Instructmeta-llama	1127	37.7K	8.0B	Llama 3	359.6K	5,787
118	gemma-2-2b-itgoogle	1117	18.2K	2.6B	Gemma	125.7K	1,607
119	granite-3.1-2b-instructibm-granite	1116	1.3K	2.5B	Apache 2.0	24.5K	57
120	Qwen1.5-32B-ChatQwen	1115	7.7K	32.5B	Tongyi Qianwen	46.8K	115
121	Phi-3-medium-4k-instructmicrosoft	1113	9.4K	14.0B	MIT	18.7K	231
122	Mixtral-8x7B-Instruct-v0.1mistralai	1109	25K	46.7B	Apache 2.0	298.7K	4,777
123	Qwen1.5-14B-ChatQwen	1102	6.2K	14.2B	Tongyi Qianwen	12.2K	118
124	Llama-3.2-3B-Instructmeta-llama	1089	3.2K	3.2B	Llama 3.2	914.1K	2,963
125	granite-3.0-8b-instructibm-granite	1088	2.6K	8.2B	Apache 2.0	13.3K	209
126	Phi-3-small-8k-instructmicrosoft	1086	6.6K	7.4B	MIT	7.1K	182
127	deepseek-llm-67b-chatdeepseek-ai	1079	1.5K	—	Custom	2.8K	222
128	Llama-2-70b-chat-hfmeta-llama	1071	12.6K	69.0B	Llama 2	4.8K	2,211
129	Mistral-7B-Instruct-v0.2mistralai	1060	6.7K	7.2B	Apache 2.0	1.6M	3,280
130	Qwen1.5-7B-ChatQwen	1058	1.7K	7.7B	Tongyi Qianwen	7K	192
131	Phi-3-mini-4k-instructmicrosoft	1058	4.4K	3.8B	MIT	329.2K	1,489
132	gemma-1.1-7b-itgoogle	1057	8.9K	8.5B	Gemma	4.2K	276
133	granite-3.0-2b-instructibm-granite	1056	2.7K	2.6B	Apache 2.0	3.2K	48
134	SOLAR-10.7B-Instruct-v1.0upstage	1046	1.2K	10.7B	CC BY-NC 4.0	48.7K	660
135	Llama-2-13b-chat-hfmeta-llama	1045	6.1K	13.0B	Llama 2	148.8K	1,138
136	Qwen-14B-ChatQwen	1031	1.5K	14.2B	—	5.2K	381
137	Llama-3.2-1B-Instructmeta-llama	1031	3.2K	1.2B	Llama 3.2	1.5M	1,961
138	Phi-3-mini-128k-instructmicrosoft	1023	7.4K	3.8B	MIT	238.7K	1,713
139	gemma-7b-itgoogle	1017	2.8K	8.5B	Gemma	25.3K	1,263
140	Llama-2-7b-chat-hfmeta-llama	1006	4.5K	6.7B	Llama 2	126.4K	4,951
141	gemma-1.1-2b-itgoogle	992	3.9K	2.5B	Gemma	6.1K	175
142	Qwen1.5-4B-ChatQwen	978	2.6K	4.0B	Custom	15.9K	53
143	OLMo-7B-Instructallenai	978	1.9K	6.9B	Apache 2.0	846	54
144	gemma-2b-itgoogle	970	1.5K	2.5B	Gemma	14K	975

Show the top 20

Scores: LMArena, CC BY 4.0. Measured on served versions, not local builds. A model with several served variants shows its best.