Providers/DeepInfra
DeepInfra logo

DeepInfra

Inference providerdeepinfra.com ↗Status ↗
MODELS
69
ENDPOINTS
136
MIN INPUT
$0.02
MEDIAN BLENDED
$0.25
AVG UPTIME
95.6%
CHEAPEST FOR
16 models
Models on DeepInfra · prices & measurements for models served by DeepInfra’s APIUSD / MTOK · VERIFIED 15 MINUTES AGO
MODELINPUTOUTPUTCONTEXTQUANTTTFTTPSSMOKEVS BEST ELSEWHERE
Meta logoLlama 3.1 8B InstructMetaBEST PRICE$0.02$0.04131Kfp80.9s°25° tok/s3/5cheapest here
Mistral logoMistral NemoMistral$0.023$0.024131Kfp80.55s°26° tok/s—$0.017 elsewhere
OpenAI logogpt-oss-20bOpenAI$0.024$0.112131Kbf16✕✕2/5$0.017 elsewhere
OpenAI logogpt-oss-120bOpenAI$0.037$0.17131Kbf160.58s°38° tok/s4/5$0.03 elsewhere
Google logoGemma 3 4BGoogle$0.04$0.08131Kbf160.8s°46° tok/s—$0.04 elsewhere
Google logoGemma 3 12BGoogle$0.04$0.12131Kbf161.6s°34° tok/s—$0.35 elsewhere
Mistral logoMistral Small 3Mistral$0.04$0.06433Kfp80.95s°32° tok/s—only provider
Sao10K logoLlama 3 8B LunarisSao10KBEST PRICE$0.04$0.058Kfp80.89s°89° tok/s—cheapest here
DeepSeek logoDeepSeek V4 Flash 0731DeepSeek$0.048$0.1441.05Mfp80.75s°79° tok/s—$0.044 elsewhere
Microsoft logoPhi 4Microsoft$0.056$0.11216Kbf160.98s°79° tok/s—only provider
InclusionAI logoLing 3.0 Flash VLInclusionAI$0.06$0.18131Kfp40.24s°219° tok/s—only provider
IBM Granite logoGranite 4.2 8BIBM GraniteBEST PRICE$0.06$0.25131Kbf160.76s°159° tok/s—cheapest here
NVIDIA logoNemotron 3.5 LightningNVIDIA$0.06$0.16262Kbf160.34s°216° tok/s—$0.02 elsewhere
InclusionAI logoLing 3.0 FlashInclusionAI$0.06$0.18131Kfp40.37s°192° tok/s—$0.021 elsewhere
NVIDIA logoNemotron 3 SuperNVIDIA$0.06$0.32262Kbf16✕✕—$0.08 elsewhere
Mistral logoMistral Small 3.2 24BMistral$0.06$0.16128Kfp80.72s°111° tok/s—$0.09 elsewhere
Qwen logoQwen3 32BQwen$0.064$0.22441Kfp81.4s°21° tok/s—$0.14 elsewhere
Google logoGemma 3 27BGoogle$0.064$0.128131Kfp80.94s°33° tok/s3/5$0.08 elsewhere
Google logoGemma 4 26B A4B Google$0.07$0.34262Kfp80.67s°58° tok/s5/5$0.05 elsewhere
DeepSeek logoDeepSeek V4 Flash 0423DeepSeek$0.072$0.1441.05Mfp80.8s°34° tok/s5/5$0.014 elsewhere
Qwen logoQwen3 Next 80B A3B InstructQwen$0.072$0.88262Kfp80.59s°172° tok/s—$0.10 elsewhere
Qwen logoQwen3 235B A22B Instruct 2507Qwen$0.072$0.44262Kfp81.2s°32° tok/s5/5$0.14 elsewhere
DeepSeek logoDeepSeek V4.1 Flash 1017DeepSeek$0.078$0.2351.05Mfp80.38s°103° tok/s—$0.036 elsewhere
Qwen logoQwen3.6 35B A3BQwen$0.08$0.76262Kfp80.47s°219° tok/s2/5$0.045 elsewhere
Qwen logoQwen3.5-9BQwen$0.08$0.12262Kbf160.71s°75° tok/s—$0.08 elsewhere
Meta logoLlama 4 ScoutMeta$0.08$0.24328Kfp80.28s°118° tok/s—only provider
Qwen logoQwen3 30B A3BQwen$0.096$0.4041Kfp80.53s°118° tok/s—only provider
Qwen logoQwen3 14BQwen$0.096$0.19241Kfp81.6s°46° tok/s—$0.10 elsewhere
Meta logoLlama 3.3 70B InstructMetaBEST PRICE$0.10$0.32131Kfp82.1s°10° tok/s5/5cheapest here
Z.ai logoGLM 5.3 FlashZ.ai$0.101$0.3351.05Mfp40.69s°87° tok/s—$0.04 elsewhere
DeepSeek logoDeepSeek V4.1 FlashDeepSeek$0.112$0.3361.05Mfp80.48s°149° tok/s—$0.05 elsewhere
Qwen logoQwen3.5-35B-A3BQwen$0.112$0.80262Kfp84.2s°62° tok/s3/5$0.065 elsewhere
Google logoGemma 4 31BGoogle$0.135$0.51262Kfp40.56s°30° tok/s5/5$0.10 elsewhere
Xiaomi logoMiMo-V2.6-FlashXiaomi$0.14$0.281.05Mfp81.1s°51° tok/s—$0.10 elsewhere
Meta logoLlama Guard 4 12BMeta$0.144$0.144164Kbf161.2s°——only provider
RADaisyRose 27BRadixark$0.15$1.88262Kbf160.87s°191° tok/s—only provider
NVIDIA logoNemotron 3.5 Super VLNVIDIA$0.15$0.50262Kfp40.76s°107° tok/s—$0.06 elsewhere
Qwen logoQwen3.8 27BQwen$0.15$1.88262Kfp80.95s°62° tok/s—$0.035 elsewhere
Qwen logoQwen3 VL 30B A3B InstructQwenBEST PRICE$0.15$0.60262Kfp80.57s°19° tok/s—cheapest here
DeepSeek logoDeepSeek V4 Flash Vision ExpDeepSeek$0.172$0.5171.05Mfp83.2s°225° tok/s—$0.44 elsewhere
NVIDIA logoNemotron 3.5 Content SafetyNVIDIA$0.20$0.20131Kbf160.26s°263° tok/s—only provider
Qwen logoQwen3 VL 235B A22B InstructQwenBEST PRICE$0.20$0.88262Kfp8✕✕—cheapest here
MiniMax logoMiniMax M3MiniMax$0.224$0.88524Kfp80.65s°40° tok/s5/5$0.23 elsewhere
DeepSeek logoDeepSeek V3.1DeepSeekBEST PRICE$0.25$0.95164Kfp41.2s°24° tok/s5/5cheapest here
Qwen logoQwen3.6 27BQwen$0.256$2.56262Kfp80.9s°83° tok/s2/5$0.25 elsewhere
DeepSeek logoDeepSeek V3DeepSeek$0.256$0.712164Kfp40.52s°53° tok/s—$0.257 elsewhere
Qwen logoQwen3.5-27BQwen$0.26$2.60262Kfp81.1s°64° tok/s3/5$0.195 elsewhere
DeepSeek logoDeepSeek V3.2DeepSeek$0.26$0.38164Kfp40.9s°31° tok/s5/5$0.259 elsewhere
Thinking Machines logoInkling 2 SmallThinking Machines$0.285$0.951.05Mfp4———$0.15 elsewhere
Qwen logoQwen2.5 72B InstructQwen$0.288$0.3233Kfp81.9s°42° tok/s—only provider
Meta logoMuse Glimmer 30BMetaBEST PRICE$0.30$1.20131Kbf161.1s°157° tok/s—cheapest here
Qwen logoQwen3 Coder 480B A35BQwenBEST PRICE$0.30$1.00262Kfp40.42s°162° tok/s—cheapest here
Z.ai logoGLM 4.7Z.ai$0.32$1.40203Kfp45.2s°26° tok/s2/5$0.40 elsewhere
Qwen logoQwen3.5 397B A17BQwen$0.36$2.40262Kfp81.2s°50° tok/s3/5$0.39 elsewhere
Xiaomi logoMiMo-V2.6-ProXiaomiBEST PRICE$0.387$0.7831.05Mfp80.73s°57° tok/s—cheapest here
NVIDIA logoNemotron 3 UltraNVIDIA$0.40$1.76262Kfp41s°32° tok/s—$0.50 elsewhere
Z.ai logoGLM 4.6Z.ai$0.40$1.60203Kfp40.69s°22° tok/s3/5$0.43 elsewhere
Z.ai logoGLM 5.3Z.ai$0.45$2.001.05Mfp40.67s°88° tok/s—$0.07 elsewhere
Thinking Machines logoInkling SmallThinking Machines$0.45$1.20524Kfp40.36s°355° tok/s—only provider
Z.ai logoGLM 5.2Z.ai$0.45$1.441.05Mfp40.86s°101° tok/s—$0.007 elsewhere
DeepSeek logoR1 0528DeepSeekBEST PRICE$0.50$2.15164Kfp40.92s°20° tok/s—cheapest here
Moonshot AI logoKimi K2.6Moonshot AI$0.60$2.80262Kfp40.81s°26° tok/s—$0.39 elsewhere
Sao10K logoLlama 3.1 Euryale 70B v2.2Sao10K$0.68$0.68131Kfp82.3s°40° tok/s—only provider
Tencent logoHy4Tencent$0.834$2.501.05Mfp4———$0.417 elsewhere
Tencent logoHy4 previewTencentBEST PRICE$0.834$2.501.05Mfp41.3s°124° tok/s—cheapest here
DeepSeek logoDeepSeek V4 Pro 0813DeepSeek$1.04$2.081.05Mfp81.2s°37° tok/s—$0.10 elsewhere
DeepSeek logoDeepSeek V4 Pro 0423DeepSeek$1.04$2.081.05Mfp80.98s°57° tok/s—$0.661 elsewhere
Moonshot AI logoKimi K3.1 ProMoonshot AI$1.90$9.501.05Mbf16✕✕—$0.24 elsewhere
Moonshot AI logoKimi K3Moonshot AI$2.85$14.251.05Mmxfp41.2s°218° tok/s4/5$0.182 elsewhere
TTFT/TPS: ModelIndex-measured medians (P50, trailing 21d) for DeepInfra’s endpoints · ° = measured via OpenRouter routing · ✕ = probed, no measurable stream · SMOKE: of 5 capability probes passed · methodology on each model page