Best open weights

Highest-quality models you can download and run yourself.

174 models qualify · data as of 9 Sept 2026

1. 1Kimi K366.5
2. 2GLM-5.3-Flash61
3. 3Kimi K2.660.6
4. 4GLM-5.160.1
5. 5DeepSeek V4 Pro60
6. 6GLM-5.359.9
7. 7DeepSeek V4 Flash59.3
8. 8Qwen3.8-Flash-Next59.3
9. 9Qwen3.6 Plus59
10. 10GLM-559
11. 11MiMo-V2.5-Pro59
12. 12Qwen3.5 397B-A17B58.6
13. 13Nemotron 3 Ultra 550B A55B58.3
14. 14MiMo-V2.558.3
15. 15GLM-5.257.6

SearchAll providersAI21 LabsAlibabaAllen Institute for AIBaiduCohereDeepSeekGoogleIBMInclusionAILG AI ResearchLiquid AILongCatMBZUAI Institute of Foundation ModelsMetaMicrosoftMiniMaxMistral AIMoonshot AIMultiverse ComputingNanbeigeNaverNous ResearchNVIDIAOpenAIOpenBMBPoolsideReka AISarvamStepFunSwiss AI InitiativeThinking MachinesTII UAETrillion LabsUpstageXiaomiZhipu AIOpen weights174 of 174

#	Model	Index ↓	Answer ↕	Price $/M ↕	Speed tok/s ↕
1	Kimi K3Moonshot AIopen ↗	66.5	66.5	$6.00	37
2	GLM-5.3-FlashZhipu AIopen ↗	61	61.0	$0.119	72
3	Kimi K2.6Moonshot AIopen ↗	60.6	60.6	$1.71	52
4	GLM-5.1Zhipu AIopen ↗	60.1	60.1	$2.15	60
5	DeepSeek V4 ProhighDeepSeekopen ↗	60	60.0	$0.548	66
6	GLM-5.3maxZhipu AIopen ↗	59.9	59.9	$2.15	60
7	DeepSeek V4 FlashhighDeepSeekopen ↗	59.3	59.3	$0.168	118
8	Qwen3.8-Flash-NextAlibabaopen ↗	59.3	59.3	$0.230	64
9	Qwen3.6 PlusAlibabaopen	59	59.0	$1.13	56
10	GLM-5Zhipu AIopen ↗	59	59.0	$1.55	66
11	XMiMo-V2.5-ProXiaomiopen ↗	59	59.0	$0.653	35
12	Qwen3.5 397B-A17BAlibabaopen ↗	58.6	58.6	$0.387	81
13	Nemotron 3 Ultra 550B A55BNVIDIAopen ↗	58.3	58.3	$1.00	156
14	XMiMo-V2.5Xiaomiopen ↗	58.3	58.3	$0.175	54
15	GLM-5.2maxZhipu AIopen ↗	57.6	57.6	$1.79	63
16	Qwen 3.8 27BAlibabaopen ↗	57.1	57.1	$1.13	46
17	MiniMax-M3MiniMaxopen ↗	56.9	56.9	$0.525	100
18	MiniMax-M2.7MiniMaxopen ↗	56.5	56.5	$0.525	62
19	Kimi K2.7 CodeMoonshot AIopen ↗	56.2	56.2	$1.71	63
20	Kimi K2.5Moonshot AIopen ↗	56.1	56.1	$1.20	37
21	Qwen3.5 122B-A10BAlibabaopen ↗	56	56.0	$1.10	129
22	DeepSeek V3.2thinkingDeepSeekopen ↗	55.9	55.9	$0.315	30
23	Gemma-4-31B-ITNVIDIAopen ↗	55.4	55.4	$0.205	35
24	Qwen3.5 27BAlibabaopen ↗	55.4	55.4	$0.825	76
25	MiniMax-M2MiniMaxopen ↗	55.1	55.1	$0.525	88
26	Muse Glimmer (high)Metaopen	55.1	55.1	$0.638	104
27	TInkling-SmallThinking Machinesopen ↗	55	55.0	$0.638	150
28	GLM-4.7Zhipu AIopen ↗	54.9	54.9	$1.00	94
29	Step 3.5 FlashStepFunopen ↗	54.9	54.9	$0.150	216
30	DeepSeek V3.1thinkingDeepSeekopen ↗	54.8	54.8	$1.92	29
31	Step 3.7 FlashStepFunopen ↗	54.6	54.6	$0.416	86
32	Gemma 4 26B A4BGoogleopen ↗	54	54.0	$0.190	30
33	ILing 3.0 FlashInclusionAIopen ↗	54	54.0	$0.115	290
34	Command A+Cohereopen	53.9	53.9	–	235
35	Nemotron 3 SuperNVIDIAopen ↗	53.6	53.6	$0.350	96
36	MiniMax-M2.5MiniMaxopen ↗	53.5	53.5	$0.525	93
37	Qwen3.5 35B-A3BAlibabaopen ↗	53.4	53.4	$0.688	145
38	XMiMo-V2-FlashthinkingXiaomiopen ↗	53.4	53.4	$0.150	–
39	Qwen3.6 27BAlibabaopen ↗	53.2	53.2	$1.35	56
40	Qwen3.6 35B-A3BAlibabaopen ↗	53.2	53.2	$0.557	126
41	Kimi K2 ThinkingthinkingMoonshot AIopen ↗	52.8	52.8	$1.07	119
42	GLM-4.6thinkingZhipu AIopen ↗	52.3	52.3	$0.963	44
43	Qwen3-VL 235B-A22BthinkingAlibabaopen ↗	52.1	52.1	$1.30	56
44	Gemma 4 12BGoogleopen ↗	52.1	52.1	$0.150	107
45	Qwen3 235B A22BthinkingAlibabaopen ↗	51.7	51.7	$0.975	60
46	Granite 4.2 30BIBMopen ↗	51.3	51.3	$0.283	76
47	Qwen3-Next 80B A3BthinkingAlibabaopen ↗	51	51.0	$0.466	190
48	Nemotron Cascade 2 30B A3BNVIDIAopen ↗	51	51.0	–	–
49	LK-EXAONELG AI Researchopen	50.9	50.9	–	–
50	gpt-oss-120bhighOpenAIopen ↗	50.5	50.5	–	202
51	GLM-4.5Zhipu AIopen ↗	50.2	50.2	$1.00	17
52	GLM-4.7-FlashZhipu AIopen ↗	50	50.0	$0.153	97
53	LEXAONE 4.5 33BLG AI Researchopen ↗	50	50.0	–	–
54	North Mini CodeCohereopen	49.8	49.8	–	44
55	MHyperNova 60B 2605 (high, based on gpt-oss-120b)Multiverse Computingopen	49.4	49.4	$0.065	345
56	MK2 Think V2thinkingMBZUAI Institute of Foundation Modelsopen	49.3	49.3	–	–
57	DeepSeek R1DeepSeekopen ↗	48.9	48.9	$0.315	19
58	Kimi K2Moonshot AIopen ↗	48.8	48.8	$1.07	40
59	Granite 4.2 8BIBMopen ↗	48.6	48.6	$0.108	120
60	nemotron-3-nano-30b-a3bthinkingNVIDIAopen ↗	48.6	48.6	$0.088	231

Show 100 more