Skip to content
ModelCap

Open-weight ranking

Best open-weight LLMs

Language models whose weights can be downloaded — openly or behind a licence gate — ordered by their position on the live ModelCap Index. The same evidence, the same uncertainty, restricted to models you can self-host.

Snapshot as of 22 September 2026

MiMo-V2.6-Pro from Xiaomi leads the open-weight LLM list as of 22 September 2026 with an Index score of 88.1, holding ModelCap Index position #5; 167 models qualify.

Models on this list
167
With openly downloadable weights
96
With a listed API price
47

Best open-weight LLMs (top 100 of 167)

Every row is a current, canonical language model with a public position on the ModelCap Index; the list filters by a published fact and orders by a published figure.

Best open-weight LLMs: current, canonical language models with a public ModelCap Index position whose weight access is open or gated on the Hugging Face Hub, ordered by ModelCap Index rank
#ModelWeightsModelCap IndexInput $/1MOutput $/1MContext
1MiMo-V2.6-ProXiaomiOpen weights#588.1$0.435$0.871M
2MiMo-V2.6-FlashXiaomiOpen weights#1084.4$0.14$0.281M
3GLM 5.3 FlashZ.aiOpen weights#1481.9$0.15$0.501.3M
4LongCat 2.0MeituanOpen weights#1681.7$0.30$1.201M
5DeepSeek V4.1 FlashDeepSeekOpen weights#1980.5$0.15$0.601M
6Nex N2.5 MaxNex AGIOpen weights#2080.4
7Nex-N2.5-ProNex AGIOpen weights#2379.7$0.075$0.25262K
8Ornith 1.5 397BOrnith AiOpen weights#2479.4
9DeepSeek V4 Pro 0813DeepSeekOpen weights#2778.5$0.66$1.981M
10Hy4 previewTencentOpen weights#2876.4$0.834$2.501M
11Gemma 4 31BGoogleOpen weights#2973.8$0.09$0.34262K
12Nex-N2.5-MiniNex AGIOpen weights#3173.2$0.025$0.10262K
13K2 Horizon 375B A23BIFMOpen weights#3272.5
14Muse Glimmer 30BMetaOpen weights#3571.4$0.30$1.20131K
15Qwen3.8 27BQwenOpen weights#3671.0$0.42$3.001M
16Gemma 4 26B A4BGoogleOpen weights#3870.2$0.09$0.30262K
17Ornith 1.5 35B A3BOrnith AiOpen weights#3970.2
18KAT Coder V2.5 DevKwaipilotOpen weights#4070.2
19Qwen3.5 397B A17BQwenOpen weights#4269.7$0.55$3.50262K
20MiMo-V2.5XiaomiOpen weights#4369.6$0.14$0.281M
21Laguna M.1PoolsideOpen weights#4469.1
22InklingThinking MachinesOpen weights#4668.6$1.00$4.051M
23Ling 3.0 Flash VLInclusionAIOpen weights#5067.0$0.06$0.18131K
24MiMo V2 FlashXiaomiMiMoOpen weights#5166.8
25gemma 4 12B itGoogleOpen weights#5266.7
26LongCat Flash Thinking 2601Meituan LongcatOpen weights#5466.4
27Step3 VL 10BStepfun AiOpen weights#5565.6
28Ornith 1.5 9BOrnith AiOpen weights#5665.4
29Xing4.0 29B A4BXingChen AGIOpen weights#5765.4
30Ling 3.0 Flash FinInclusionAIOpen weights#5865.1$0.06$0.18262K
31Qwen3 235B A22B Instruct 2507QwenOpen weights#6063.6$0.087$0.35262K
32DeepSeek V3.2DeepSeekOpen weights#6163.4$0.269$0.40164K
33R1 0528DeepSeekOpen weights#6263.1$0.50$2.15164K
34Qwen SEA LION v4.5 27B ITAisingaporeOpen weights#6561.7
35Qwen Image BenchQwenOpen weights#6661.7
36UI Mate 27BTencentOpen weights#6761.7
37UI Mate democua 27BTencentOpen weights#6861.7
38Qwen3 VL 235B A22B InstructQwenOpen weights#6961.4$0.21$1.90262K
39DeepSeek V3.1 TerminusDeepSeekOpen weights#7161.1$0.27$1.00164K
40Qwen3.5-122B-A10BQwenOpen weights#7360.8$0.26$2.08262K
41Qwen3.6 35B A3BQwenOpen weights#7460.8$0.15$1.00262K
42Intern S2 PreviewInternlmOpen weights#7560.2
43LongCat Flash ChatMeituan LongcatOpen weights#7659.6
44Ring 2.6 1TInclusionAIOpen weights#7759.3
45Ling 3.0 FlashInclusionAIOpen weights#7859.0$0.021$0.063262K
46K2 Horizon MoVA 36B A4BIFMOpen weights#7958.9
47Laguna XS.2PoolsideOpen weights#8257.0
48Fara1.5 27BMicrosoftOpen weights#8356.8
49Inkling SmallThinking MachinesOpen weights#8456.6$0.45$1.201M
50MiniMax M1 80kMiniMaxAIOpen weights#8556.51M
51Qwen3 235B A22BQwenOpen weights#8656.0$0.455$1.82131K
52Agent.XortronDarkc0deOpen weights#8855.6
53Qwen3 Next 80B A3B InstructQwenOpen weights#8955.4$0.09$1.10262K
54Qwen3 235B A22B Thinking 2507QwenOpen weights#9055.4$0.23$2.30131K
55MiniCPM5 2BOpenbmbOpen weights#9155.2
56Qwen3 Coder 30B A3B InstructQwenOpen weights#9254.8$0.07$0.28262K
57Qwen3 VL 235B A22B ThinkingQwenOpen weights#9354.6$0.40$4.00131K
58AReaL SEA 235B A22BInclusionAIOpen weights#9554.2
59Qwen3 Nemotron 235B A22B GenRM 2603NVIDIAOpen weights#9654.2
60Step 3.7 FlashStepFunOpen weights#9754.2$0.20$1.15262K
61Mistral Small 4Mistral AIOpen weights#9854.1$0.15$0.60262K
62MAI DS R1MicrosoftOpen weights#10053.3
63K2 Horizon 32BIFMOpen weights#10253.1
64North Mini CodeCohereOpen weights#10352.8FreeFree256K
65Qwen3 Coder 480B A35BQwenOpen weights#10452.3$0.30$1.00262K
66Qwen3 Coder NextQwenOpen weights#10552.1$0.12$0.80262K
67Intern S2 Preview 397BInternlmOpen weights#10651.8
68Qwen3 30B A3B Instruct 2507QwenOpen weights#10850.5$0.048$0.193262K
69Nanbeige4.1 3BNanbeigeOpen weights#10949.3
70GLM 4.6VZ.aiOpen weights#11048.5$0.30$0.90131K
71gemma 4 E4B itGoogleOpen weights#11248.2
72GLM 4.5 AirZ.aiOpen weights#11348.1$0.13$0.85131K
73Qwen3.5-9BQwenOpen weights#11547.2$0.10$0.15262K
74Qwen3 Next 80B A3B ThinkingQwenOpen weights#11746.7$0.15$1.20262K
75gemma 4 E2B itGoogleOpen weights#11846.1
76Fara1.5 9BMicrosoftOpen weights#12046.0
77UI Mate 9BTencentOpen weights#12146.0
78MiMo V2.6 Distill Qwen 9BXiaomiMiMoOpen weights#12246.0
79Qwen2.5 VL 32B InstructQwenOpen weights#12345.8
80Gemma SEA LION v4.5 E2B ITAisingaporeOpen weights#12444.9
81Qwen3.5 4BQwenOpen weights#12544.6
82Phi mini MoE instructMicrosoftOpen weights#12744.2
83Gemma 3 27BGoogleGated access · Licence-restricted · gemma#12843.7$0.08$0.45131K
84Fara1.5 4BMicrosoftOpen weights#13043.4
85Qwen Drive 1.0 4BQwenOpen weights#13143.4
86gpt-oss-120bOpenAIOpen weights#13243.3$0.15$0.60131K
87Mistral Small 3.2 24BMistral AIOpen weights#13441.9$0.094$0.25256K
88Command ACohereGated access · Licence-restricted · cc-by-nc-4.0#13740.5$2.50$10.00256K
89Qwen3 32BQwenOpen weights#14039.8$0.08$0.28131K
90gpt oss safeguard 120bOpenAIOpen weights#14239.3
91command a reasoning 08 2025CohereLabsGated access · Licence-restricted · cc-by-nc-4.0#14339.3
92command a translate 08 2025CohereLabsGated access · Licence-restricted · cc-by-nc-4.0#14439.3
93Qwen SEA LION v4 32B ITAisingaporeOpen weights#14638.6
94GDN primed HQwen3 32B InstructAmazonOpen weights#14738.6
95GKA primed HQwen3 32B InstructAmazonOpen weights#14838.6
96GKA primed HQwen3 32B ReasonerAmazonOpen weights#14938.6
97FrogBoss 32B 2510MicrosoftOpen weights#15038.6
98granite 4.2 30bIBM GraniteOpen weights#15338.5131K
99Qwen3 4B Instruct 2507QwenOpen weights#15437.5262K
100QwQ 32BQwenOpen weights#15536.8

Best open-weight LLMs: common questions

Which model tops the open-weight LLM list right now?

MiMo-V2.6-Pro (Xiaomi), as of 22 September 2026, with an Index score of 88.1 and ModelCap Index position #5.

How is the open-weight LLM list built?

It contains current, canonical language models with a public ModelCap Index position whose weight access is open or gated on the Hugging Face Hub, ordered by ModelCap Index rank. Nothing is estimated for the list itself: the filter is a published fact and the order is a published figure, and every row links to the model page that shows its evidence.

How many models qualify for the open-weight LLM list?

167 models qualify; the page shows the first 100. The population is the live ModelCap Index, so a model appears here the moment it holds a public position and meets the filter.

What is the cheapest model on the open-weight LLM list?

Ling 3.0 Flash lists the lowest blended API price at $0.032 per 1M tokens ($0.021 input / $0.063 output).

How current is the open-weight LLM list?

It is rendered from the sealed dataset published 22 September 2026 and re-renders within a minute of each refresh; positions, prices and context figures are the ones shown on the live rankings at the same instant.

Explore further