Models with downloadable open weights — runnable on your own infrastructure. Not a hosted inference provider. Self-hosted · Open Weights · Self Hosted

Intelligence vs Price

Best value among Self-hosted models on this chart: GLM-5.2 · MiniMax M3 · DeepSeek V4 Flash (and 4 more on the dashed frontier). Hover any dot for full pricing, or click a creator in the legend to isolate.

Language Models
Intelligence
Blended Price, $
Log X

Self-hosted models

165 models, 140 with pricing
All Model Types
All Creators
US Dollar ($)
Per 1M tokens
Input/1M
to
Output/1M
to
Model
Creator
Input Price, $
Output Price, $
Context
Max Output
Inference Providers
Intelligence
Coding
GLM-5.2Zhipu AI logoZhipu AI0.8042.531.0M131Kcompare (5)51.1#168.8#1
MiniMax M3MiniMax logoMiniMax0.31.201.0M512Kcompare (5)44.4#258.6#6
DeepSeek V4 ProDeepSeek logoDeepSeek0.4350.871.0M384Kcompare (8)44.3#359.4#5
Kimi K2.6Moonshot AI (Kimi) logoMoonshot AI (Kimi)0.6843.40262K262Kcompare (8)44.2#461.8#2
MiMo V2.5 ProXiaomi logoXiaomi0.4350.871.1M131Kcompare (2)42.2#560.2#4
Kimi K2.7 CodeMoonshot AI (Kimi) logoMoonshot AI (Kimi)0.823.71262K33Kcompare (5)41.9#660.8#3
InklingThinkingmachines1.004.05524KN/Acompare (3)40.7#752.1#11
DeepSeek V4 FlashDeepSeek logoDeepSeek0.0980.1961.0M384Kcompare (11)40.3#856.2#7
GLM-5.1Zhipu AI logoZhipu AI0.9663.04205K131Kcompare (6)40.2#955.8#8
GLM-5Zhipu AI logoZhipu AI0.5732.55205K131Kcompare (9)39.5#10N/A
MiniMax M2.7MiniMax logoMiniMax0.250.551.0M131Kcompare (6)38.1#1152.6#10
Nemotron 3 Ultra 550B A55BNVIDIA logoNVIDIA0.62.401.0M66Kcompare (3)37.8#1249.3#13
Qwen3.6 27BAlibaba logoAlibaba0.150.5262K66Kcompare (5)37.1#1353.7#9
Kimi K2.5Moonshot AI (Kimi) logoMoonshot AI (Kimi)0.52.80262K98Kcompare (12)35.4#1446.8#16
Qwen3.5 27BAlibaba logoAlibaba0.262.40262K82Kcompare (3)33.8#15N/A
Qwen3.5 397B A17BAlibaba logoAlibaba0.392.34262K66Kcompare (5)33.7#1648.2#14
MiniMax M2.5MiniMax logoMiniMax0.150.91.0M197Kcompare (8)33.7#17N/A
GLM-4.7Zhipu AI logoZhipu AI0.41.75205K131Kcompare (13)33.7#1845.3#18
Kimi K2 ThinkingMoonshot AI (Kimi) logoMoonshot AI (Kimi)0.471.20262K100Kcompare (13)32.7#19N/A
Qwen3.5 122B A10BAlibaba logoAlibaba0.251.75262K66Kcompare (4)32.3#2045.7#17
Qwen3.6 35B A3BAlibaba logoAlibaba0.140.45262K16Kcompare (6)31.6#2141.9#19
MiniMax M2.1MiniMax logoMiniMax0.31.201.0M131Kcompare (8)31.4#22N/A
Step 3.7 FlashStepFun logoStepFun0.21.15262K256Kcompare (2)30.3#2339.6#20
Qwen3.5 35B A3BAlibaba logoAlibaba0.141.00262K66Kcompare (3)29.3#24N/A
MiniMax M2MiniMax logoMiniMax0.31.20205K131Kcompare (9)28.3#25N/A
InclusionAI Ling 2.6 1TInclusionai0.0750.625262K33Kcompare (2)26.1#26N/A
Step 3.5 FlashStepFun logoStepFun0.090.3262K66Kcompare (2)25.5#27N/A
MiMo V2 FlashXiaomi logoXiaomi0.10.3262K32Kcompare (2)24.7#2949.8#12
DeepSeek V3.2DeepSeek logoDeepSeek0.2690.4164K66Kcompare (13)24.7#28N/A
GPT OSS 120BOpenAI logoOpenAI0.0370.17131K131Kcompare (23)23.8#3030.4#22
GLM-4.6Zhipu AI logoZhipu AI0.4312.00205K131Kcompare (10)23.0#31N/A
GLM-4.7 FlashZhipu AI logoZhipu AI0.06050.4203K131Kcompare (4)22.9#32N/A
DeepSeek V3.2 SpecialeDeepSeek logoDeepSeek0.581.68164KN/Acompare (1)22.2#33N/A
Qwen3.5 9BAlibaba logoAlibaba0.10.15262K33Kcompare (2)21.4#3428.7#23
DeepSeek V3.1 TerminusDeepSeek logoDeepSeek0.271.00164K66Kcompare (6)21.4#35N/A
Qwen3 Coder NextAlibaba logoAlibaba0.110.8262K8Kcompare (5)21.1#3636.2#21
DeepSeek V3.1DeepSeek logoDeepSeek0.250.95164K128Kcompare (14)21.0#37N/A
GLM-4.5Zhipu AI logoZhipu AI0.41.60131K98Kcompare (7)19.5#38N/A
Kimi K2Moonshot AI (Kimi) logoMoonshot AI (Kimi)0.572.30262K100Kcompare (4)19.4#3947.1#15
DeepSeek R1DeepSeek logoDeepSeek0.280.4164K66Kcompare (14)18.5#4024.6#25
Arcee Trinity Large ThinkingArcee AI logoArcee AI0.250.8262K80Kcompare (2)18.2#4125.8#24
Qwen3 Coder 480B A35B InstructAlibaba logoAlibaba0.221.30262K66Kcompare (8)18.0#42N/A
MiniMax M1 80kMiniMax logoMiniMax0.10.11.0M40Kcompare (3)17.7#43N/A
GLM-4.5 AirZhipu AI logoZhipu AI0.1250.45131K98Kcompare (7)16.5#44N/A
DeepSeek V3 324DeepSeek logoDeepSeek0.20.4164K16Kcompare (13)15.4#4521.2#27
GPT OSS 20BOpenAI logoOpenAI0.01450.07131K131Kcompare (18)14.9#4620.7#28
Qwen3 VL 235B A22B InstructAlibaba logoAlibaba0.210.88262K129Kcompare (7)14.3#47N/A
Llama 4 MaverickMeta logoMeta0.120.4851.0M16Kcompare (6)14.3#4816.3#29
DeepSeek V3DeepSeek logoDeepSeek0.20.2164K82Kcompare (12)14.2#4923.0#26
Qwen3 Next 80B A3B InstructAlibaba logoAlibaba0.10.9262K66Kcompare (10)13.7#50N/A
Qwen3 Coder 30B A3B InstructAlibaba logoAlibaba0.070.27262K33Kcompare (5)13.6#51N/A
QwQ 32BAlibaba logoAlibaba0.150.2131K16Kcompare (7)13.4#52N/A
Qwen3 VL 32B InstructAlibaba logoAlibaba0.1040.416131K33Kcompare (3)11.1#53N/A
GLM-4.6VZhipu AI logoZhipu AI0.30.9131K33Kcompare (4)11.0#55N/A
DeepSeek R1 Distill Qwen 32BDeepSeek logoDeepSeek0.150.15131K32Kcompare (6)11.0#54N/A
Qwen3 235B A22B InstructAlibaba logoAlibaba0.090.58262K16Kcompare (11)10.9#56N/A
Qwen3 VL 30B A3B InstructAlibaba logoAlibaba0.130.52262K33Kcompare (5)10.0#57N/A
Llama 4 ScoutMeta logoMeta0.10.31.3M16Kcompare (4)10.0#588.2#31
DeepSeek R1 Distill Llama 70BDeepSeek logoDeepSeek0.20.375131K8Kcompare (11)9.9#59N/A
Qwen2.5 72B InstructAlibaba logoAlibaba0.120.3131K16Kcompare (7)9.6#60N/A
Llama 3.3 70B InstructMeta logoMeta0.120.2131K128Kcompare (21)9.4#6111.9#30
Qwen3 30B A3BAlibaba logoAlibaba0.05090.2131K20Kcompare (8)9.1#62N/A
Qwen3 VL 8B InstructAlibaba logoAlibaba0.080.2262K33Kcompare (4)8.4#63N/A
Llama 3.1 Nemotron 70B InstructNVIDIA logoNVIDIA0.60.6131K16Kcompare (2)7.6#65N/A
Llama 3.1 8B InstructMeta logoMeta0.020.03200K128Kcompare (20)7.6#645.4#32
Nemotron Nano 3 30B A3BNVIDIA logoNVIDIA0.050.2262K228Kcompare (2)7.4#66N/A
Qwen2.5 Coder 32B InstructAlibaba logoAlibaba0.050.1131K4Kcompare (8)7.1#67N/A
GLM-4.5VZhipu AI logoZhipu AI0.61.20131K32Kcompare (6)7.0#68N/A
Qwen3 30B A3B InstructAlibaba logoAlibaba0.10.3262K32Kcompare (3)6.8#69N/A
Llama 3.1 70B InstructMeta logoMeta0.120.3131K16Kcompare (13)6.8#70N/A
Hermes 3 Llama 3.1 70BNous Research logoNous Research0.120.3131K16Kcompare (3)5.1#71N/A
Phi-4Microsoft logoMicrosoft0.070.1416K16Kcompare (3)4.9#72N/A
Mixtral 8x22B InstructMistral AI logoMistral AI0.60.666K2Kcompare (6)4.4#73N/A
Llama 3.2 3B InstructMeta logoMeta0.0150.02131K80Kcompare (10)4.2#74N/A
Llama 3 70B InstructMeta logoMeta0.120.3131K8Kcompare (8)3.5#75N/A
Mistral 7B InstructMistral AI logoMistral AI0.010.1127K16Kcompare (8)2.1#76N/A
Llama 3 8B InstructMeta logoMeta0.030.0432K8Kcompare (7)1.2#77N/A
Llama 3.2 1B InstructMeta logoMeta0.0270.08128K16Kcompare (6)1.1#78N/A
Arctic 2 Embed LSnowflake logoSnowflake0.10.18KN/Acompare (1)N/AN/A
AutoGLM Phone 9B MultilingualZhipu AI logoZhipu AI0.0350.13866KN/Acompare (2)N/AN/A
BGE Base EN V1.5Baai0.008N/A154KN/Acompare (2)N/AN/A
BGE Large EN V1.5Baai0.204N/A512N/Acompare (1)N/AN/A
BGE-M3Baai0.010.0160KN/Acompare (3)N/AN/A
Code Llama 7B InstructMeta logoMeta0.20.216K16Kcompare (1)N/AN/A
DeepSeek R1 528DeepSeek logoDeepSeek0.20.25164K33Kcompare (13)N/AN/A
DeepSeek V3 0324DeepSeek logoDeepSeek0.271.12164K66Kcompare (1)N/AN/A
DeepSeek V3.1DeepSeek logoDeepSeek0.250.95164K33Kcompare (1)N/AN/A
DeepSeek V3.2 ExpDeepSeek logoDeepSeek0.270.41164K66Kcompare (4)N/AN/A
DeepSeek-OCRDeepSeek logoDeepSeek0.030.038K8Kcompare (4)N/AN/A
E5 Base V2Microsoft logoMicrosoft0.060.06N/AN/Acompare (1)N/AN/A
ERNIE 4.5 21B A3BBaidu logoBaidu0.070.28120K8Kcompare (1)N/AN/A
ERNIE 4.5 VL 28B A3BBaidu logoBaidu0.140.5630K8Kcompare (1)N/AN/A
FLUX Klein 2 9BBlack Forest Labs logoBlack Forest LabsN/AN/AN/AN/Acompare (1)N/AN/A
FLUX.1 DevBlack Forest Labs logoBlack Forest Labs0.10.14K4Kcompare (1)N/AN/A
FLUX.1 SchnellBlack Forest Labs logoBlack Forest Labs0.10.14K4Kcompare (2)N/AN/A
Gemma 2 27B InstructGoogle logoGoogle0.350.658K8Kcompare (2)N/AN/A
Gemma 3 12B InstructGoogle logoGoogle0.050.1131K16Kcompare (5)N/AN/A
Gemma 3 27B InstructGoogle logoGoogle0.060.16262K16Kcompare (8)N/AN/A
Gemma 3 4B InstructGoogle logoGoogle0.040.08131K16Kcompare (3)N/AN/A
Gemma 3N E4B ITGoogle logoGoogle0.060.1233K2Kcompare (2)N/AN/A
Gemma 4 26B A4B ITGoogle logoGoogle0.070.3262K131Kcompare (6)N/AN/A
Gemma 4 31B ITGoogle logoGoogle0.120.37262K131Kcompare (6)N/AN/A
GPT OSS 20B SafeguardOpenAI logoOpenAI0.070.2131K66Kcompare (5)N/AN/A
GTE BaseAlibaba logoAlibaba0.008N/A512N/Acompare (1)N/AN/A
GTE LargeAlibaba logoAlibaba0.016N/A512N/Acompare (1)N/AN/A
Hermes 2 Pro Llama 3 8BNous Research logoNous Research0.140.148K8Kcompare (1)N/AN/A
Hermes 4 70BNous Research logoNous Research0.130.4131KN/Acompare (1)N/AN/A
Hermes Llama2 13BNous Research logoNous Research0.20.24K4Kcompare (1)N/AN/A
Kimi K2 InstructMoonshot AI (Kimi) logoMoonshot AI (Kimi)0.52.00262K33Kcompare (9)N/AN/A
L3 70B Euryale V2.1Sao10K1.481.488K8Kcompare (1)N/AN/A
L3 Lunaris 8BSao10K0.040.058K8Kcompare (3)N/AN/A
L3 Stheno 3.2 8BSao10K0.050.058KN/Acompare (2)N/AN/A
L3.3 70B Euryale V2.3Sao10K0.650.75131KN/Acompare (1)N/AN/A
L3.3 Euryale 70BSao10K0.650.75131K16Kcompare (1)N/AN/A
Llama 3.3 Nemotron 1.5 Super 49BNVIDIA logoNVIDIA0.10.4131K16Kcompare (1)N/AN/A
Llama 4 17B Maverick InstructMeta logoMeta0.050.11.0M16Kcompare (9)N/AN/A
Llama 4 17B Scout InstructMeta logoMeta0.050.110.0M16Kcompare (12)N/AN/A
LlamaGuard 3 8BMeta logoMeta0.020.03131K16Kcompare (5)N/AN/A
LlamaGuard 4 12BMeta logoMeta0.180.181.0M16Kcompare (4)N/AN/A
Magnum V4 72BAnthracite3.005.0016K2Kcompare (1)N/AN/A
Mistral 7BMistral AI logoMistral AI0.050.133K8Kcompare (5)N/AN/A
Mistral Small 3.1 24B InstructMistral AI logoMistral AI0.10.3131KN/Acompare (3)N/AN/A
Multilingual E5 LargeMicrosoft logoMicrosoft0.10.1N/AN/Acompare (1)N/AN/A
Multilingual E5 Large InstructMicrosoft logoMicrosoft0.025N/AN/AN/Acompare (2)N/AN/A
MythoMax L2 13BGryphe0.060.068K4Kcompare (4)N/AN/A
Phi-4 Mini InstructMicrosoft logoMicrosoft0.0750.3131K4Kcompare (2)N/AN/A
Qwen2.5 7B InstructAlibaba logoAlibaba0.040.0733K33Kcompare (4)N/AN/A
Qwen2.5 VL 72B InstructAlibaba logoAlibaba0.130.4131K128Kcompare (6)N/AN/A
Qwen3 14BAlibaba logoAlibaba0.060.2131K16Kcompare (7)N/AN/A
Qwen3 235B A22BAlibaba logoAlibaba0.090.54262K131Kcompare (9)N/AN/A
Qwen3 235B A22B ThinkingAlibaba logoAlibaba0.220.88262K33Kcompare (9)N/AN/A
Qwen3 32BAlibaba logoAlibaba0.050.1131K41Kcompare (15)N/AN/A
Qwen3 8BAlibaba logoAlibaba0.0350.138131K20Kcompare (6)N/AN/A
Qwen3 CoderAlibaba logoAlibaba0.31.00262K67Kcompare (2)N/AN/A
Qwen3 Embedding 8BAlibaba logoAlibaba0.02N/A41K4Kcompare (4)N/AN/A
Qwen3 Next 80B A3B ThinkingAlibaba logoAlibaba0.09750.78262K66Kcompare (9)N/AN/A
Qwen3 VL 235B A22B ThinkingAlibaba logoAlibaba0.220.88262K33Kcompare (5)N/AN/A
Qwen3 VL 30B A3B ThinkingAlibaba logoAlibaba0.130.6262K33Kcompare (4)N/AN/A
Qwen3 VL 8B ThinkingAlibaba logoAlibaba0.1171.36131K33Kcompare (2)N/AN/A
ReMM SLERP L2 13BUndi950.450.656K4Kcompare (1)N/AN/A
UI-TARS 1.5 7BByteDance logoByteDance0.10.2131K2Kcompare (1)N/AN/A
Whisper 3 LargeOpenAI logoOpenAIN/AN/AN/AN/Acompare (1)N/AN/A
Whisper 3 Large TurboOpenAI logoOpenAIN/AN/AN/AN/Acompare (3)N/AN/A