opencheap.si

Status

How every provider behind the popular models is doing, from our own requests and the test requests we send around the clock. Updated every minute, last at 18:37 UTC.

13 routes degraded
636 routes across 39 models

Degraded now

  • gpt-oss-120bGoogleDownsince Nov 8, 01:35 UTC
  • DeepSeek V4 FlashFireworksDownsince 07:33 UTC
  • DeepSeek V4.1 FlashDeepInfra · Route 3 · fp8Downsince 08:33 UTC
  • GLM 5.2NebiusDownsince 17:34 UTC
  • Kimi K3Fireworks · Route 3Downsince 17:49 UTC
  • DeepSeek V3.2BaiduDownsince 17:53 UTC
  • DeepSeek V4 FlashBaiduDownsince 17:53 UTC
  • DeepSeek V4 FlashGMICloudDownsince 17:58 UTC
  • Qwen 3 Coder 480BDeepInfraDownsince 18:22 UTC
  • Qwen/Qwen3-32BTenstorrent86.6 % uptime in the last 30 minutessince 18:31 UTC
  • GLM 5.3 FlashCoreWeave80 % uptime in the last 30 minutessince 18:33 UTC
  • DeepSeek V3.2DigitalOcean86.6 % uptime in the last 30 minutessince 18:34 UTC
  • Grok 4.7xAI · Route 283.3 % uptime in the last 30 minutessince 18:34 UTC

By model

Kimi K31 of 35 degraded
ProviderNow24 h7 d
  • AiHubMixServing100 %*100 %*
  • AlibabaServing98 %†—
  • Amazon BedrockServing98.1 %†—
  • BaseTenServing96.8 %†—
  • ChutesServing98.7 %†—
  • DecartServing100 %†—
  • DeepInfraServing99.6 %†—
  • DigitalOceanServing100 %†—
  • Fireworks · Route 1Serving92.3 %†—
  • Fireworks · Route 2Serving99.8 %†—
  • Fireworks · Route 3Down78.6 %†—
  • Hugging Face · deepinfraServing100 %*100 %*
  • InferenceNet · Route 1 · fp4Serving100 %†—
  • InferenceNet · Route 2 · fp4Serving100 %†—
  • InferenceNet · Route 3 · fp4Serving99.3 %†—
  • InferenceNet · Route 4 · fp4Serving99.6 %†—
  • Io NetServing99.7 %†—
  • MakoraServing99.9 %†—
  • ModalServing100 %†—
  • Moonshot AIServing100 %†—
  • NanoGPTServing100 %*98.3 %*
  • Parasail · Route 1 · fp4Serving100 %†—
  • Parasail · Route 2 · fp4Serving100 %†—
  • ParetoAlphaServing100 %†—
  • PhalaServing100 %†—
  • QubaxServing100 %*98.8 %*
  • RelaceServing100 %†—
  • Sail Research · Route 1 · fp4Serving100 %†—
  • Sail Research · Route 2 · fp4Serving100 %†—
  • SandbaseServing100 %*100 %*
  • SurplusServing100 %*91.4 %*
  • TogetherServing99.7 %†—
  • Wafer · Route 1 · mxfp4Serving100 %†—
  • Wafer · Route 2 · mxfp4Serving100 %†—
  • Wafer · Route 3 · mxfp4Serving100 %†—
Kimi K3.1 ProAll 29 serving
ProviderNow24 h7 d
  • AiHubMixServing100 %*100 %*
  • AkashMLServing98.7 %†—
  • AlibabaServing93.7 %†—
  • DeepInfraServing98.9 %†—
  • DigitalOceanServing100 %†—
  • FireworksServing100 %†—
  • InferenceNet · Route 1 · mxfp4Serving100 %†—
  • InferenceNet · Route 2 · mxfp4Serving100 %†—
  • InferenceNet · Route 3 · mxfp4Serving100 %†—
  • InferenceNet · Route 4 · mxfp4Serving100 %†—
  • ModalServing100 %†—
  • Moonshot AI · Route 1 · mxfp4Serving100 %†—
  • Moonshot AI · Route 2 · mxfp4Serving100 %†—
  • MorphServing100 %†—
  • NanoGPTServing100 %*100 %*
  • NebiusServing95.6 %†—
  • ParasailServing100 %†—
  • Phala · Route 1Serving100 %†—
  • Phala · Route 2Serving100 %†—
  • PrimaLabsServing99.7 %†—
  • QubaxServing100 %*98.6 %*
  • RelaceServing100 %†—
  • Sail Research · Route 1 · fp4Serving99.5 %†—
  • Sail Research · Route 2 · fp4Serving100 %†—
  • SandbaseServing100 %*100 %*
  • SurplusServing98.3 %*88.8 %*
  • TogetherServing100 %†—
  • Wafer · Route 1 · mxfp4Serving100 %†—
  • Wafer · Route 2 · mxfp4Serving99.9 %†—
Kimi K2.7 CodeAll 16 serving
ProviderNow24 h7 d
  • AiHubMixServing100 %*97.7 %*
  • AlibabaServing100 %†—
  • CloudflareServing100 %†—
  • CoreWeaveServing100 %†—
  • GMICloudServing100 %†—
  • Hugging Face · deepinfraServing100 %*100 %*
  • InceptronServing100 %†—
  • Moonshot AI · Route 1Serving100 %†—
  • Moonshot AI · Route 2Serving100 %†—
  • NanoGPTServing100 %*100 %*
  • NovitaServing100 %†—
  • QubaxServing100 %*100 %*
  • SiliconFlowServing100 %†—
  • StreamLakeServing100 %†—
  • SurplusServing98.5 %*91.1 %*
  • VeniceServing99.8 %†—
Kimi K2.6All 22 serving
ProviderNow24 h7 d
  • AiHubMixServing100 %*100 %*
  • AtlasCloudServing100 %†—
  • BaiduServing100 %†—
  • ChutesServing100 %†—
  • CloudflareServing100 %†—
  • CoreWeaveServing99.9 %†—
  • CrusoeServing99.4 %†—
  • DeepInfraServing97.7 %†—
  • DigitalOceanServing100 %†—
  • Hugging Face · novitaServing—100 %*
  • InceptronServing100 %†—
  • Moonshot AIServing99.1 %†—
  • NanoGPTServing100 %*100 %*
  • NovitaServing99.1 %†—
  • ParasailServing99.9 %†—
  • PhalaServing——
  • QubaxServing100 %*99.4 %*
  • SandbaseServing100 %*100 %*
  • SiliconFlowServing100 %†—
  • StreamLakeServing97.8 %†—
  • SurplusServing95.5 %*90.4 %*
  • VeniceServing98.9 %†—
Kimi K2.5All 11 serving
ProviderNow24 h7 d
  • AiHubMixServing100 %*94.9 %*
  • Amazon BedrockServing97.5 %†—
  • AtlasCloudServing99.1 %†—
  • Hugging Face · novitaServing100 %*100 %*
  • NanoGPTServing100 %*100 %*
  • NovitaServing98.8 %†—
  • QubaxServing100 %*100 %*
  • SandbaseServing100 %*100 %*
  • SiliconFlowServing100 %†—
  • SurplusServing97 %*93.3 %*
  • VeniceServing99.2 %†—
Kimi K2 ThinkingAll 4 serving
ProviderNow24 h7 d
  • NanoGPTServing98.5 %*99.4 %*
  • NovitaServing100 %†—
  • QubaxServing100 %*100 %*
  • SurplusServing——
GLM 5.3All 53 serving
ProviderNow24 h7 d
  • AiHubMixServing98.5 %*98.8 %*
  • AkashMLServing99.3 %†—
  • Alibaba · Route 1Serving99.1 %†—
  • Alibaba · Route 2Serving99.2 %†—
  • Amazon BedrockServing95.8 %†—
  • AtlasCloudServing100 %†—
  • BaiduServing100 %†—
  • BaseTen · Route 1 · fp4Serving99.6 %†—
  • BaseTen · Route 2 · fp4Serving100 %†—
  • BaseTen · Route 3 · fp8Serving98.5 %†—
  • BaseTen · Route 4 · fp8Serving98.2 %†—
  • CloudflareServing100 %†—
  • Community sellerServing100 %*98.3 %*
  • CoreWeaveServing100 %†—
  • CrusoeServing99.7 %†—
  • DecartServing100 %†—
  • DeepInfraServing99.5 %†—
  • DekaLLMServing99.1 %†—
  • DigitalOceanServing100 %†—
  • Fireworks · Route 1Serving97.7 %†—
  • Fireworks · Route 2Serving99.6 %†—
  • Fireworks · Route 3Serving99.6 %†—
  • FriendliServing100 %†—
  • GMICloudServing99.7 %†—
  • Hugging Face · deepinfraServing98.5 %*98.3 %*
  • InceptronServing99.7 %†—
  • InferenceNet · Route 1 · fp4Serving100 %†—
  • InferenceNet · Route 2 · fp4Serving100 %†—
  • MakoraServing99.9 %†—
  • Mistral · Route 1 · nvfp4Serving100 %†—
  • Mistral · Route 2 · nvfp4Serving99.9 %†—
  • ModalServing100 %†—
  • MorphServing96.8 %†—
  • NanoGPTServing100 %*100 %*
  • NebiusServing100 %†—
  • NovitaServing100 %†—
  • Parasail · Route 1 · fp8Serving99.8 %†—
  • Parasail · Route 2 · fp8Serving99.5 %†—
  • PhalaServing100 %†—
  • PrimeIntellectServing100 %†—
  • QubaxServing100 %*98.6 %*
  • RekaServing98.8 %†—
  • RelaceServing100 %†—
  • Sail Research · Route 1 · fp8Serving100 %†—
  • Sail Research · Route 2 · fp8Serving100 %†—
  • SandbaseServing98.4 %*99.4 %*
  • SiliconFlowServing100 %†—
  • SurplusServing80.9 %*100 %
  • TogetherServing100 %†—
  • VeniceServing100 %†—
  • Wafer · Route 1 · fp4Serving100 %†—
  • Wafer · Route 2 · fp4Serving100 %†—
  • Z.AIServing100 %†—
GLM 5.3 Flash1 of 49 degraded
ProviderNow24 h7 d
  • AiHubMixServing100 %*100 %*
  • AtlasCloudServing100 %†—
  • BaseTen · Route 1 · fp8Serving100 %†—
  • BaseTen · Route 2 · fp8Serving——
  • CohereServing100 %†—
  • Community sellerServing94.8 %*88.5 %*
  • CoreWeaveDegraded97.9 %†—
  • CrusoeServing100 %†—
  • DarkbloomServing100 %†—
  • DeepInfraServing100 %†—
  • DigitalOceanServing100 %†—
  • FireworksServing99.8 %†—
  • FriendliServing100 %†—
  • GMICloudServing99.3 %†—
  • Hugging Face · novitaServing100 %*99.4 %*
  • InceptronServing100 %†—
  • InferenceNet · Route 1 · fp4Serving100 %†—
  • InferenceNet · Route 2 · fp4Serving100 %†—
  • InferHubServing——
  • MakoraServing100 %†—
  • MistralServing100 %†—
  • ModalServing100 %†—
  • MorphServing100 %†—
  • NanoGPTServing100 %*100 %*
  • Near AIServing99.9 %†—
  • NebiusServing100 %†—
  • NextBitServing100 %†—
  • NovitaServing99.8 %†—
  • OpenInferenceServing100 %†—
  • Parasail · Route 1 · fp4Serving100 %†—
  • Parasail · Route 2 · fp4Serving99.9 %†—
  • Particle AIServing100 %†—
  • PhalaServing99.8 %†—
  • PrimeIntellectServing99.5 %†—
  • QubaxServing98.7 %*98.5 %*
  • RekaServing100 %†—
  • RelaceServing100 %†—
  • Sail Research · Route 1 · fp4Serving100 %†—
  • Sail Research · Route 2 · fp4Serving100 %†—
  • SandbaseServing100 %*98.8 %*
  • SiliconFlowServing99.8 %†—
  • StreamLakeServing99.8 %†—
  • SurplusServing88.6 %*66.6 %
  • TenstorrentServing100 %†—
  • TogetherServing100 %†—
  • VeniceServing100 %†—
  • Wafer · Route 1 · nvfp4Serving100 %†—
  • Wafer · Route 2 · nvfp4Serving100 %†—
  • Z.AIServing100 %†—
GLM 5.21 of 40 degraded
ProviderNow24 h7 d
  • AiHubMixServing100 %*99.4 %*
  • Alibaba · Route 1 · fp8Serving97.9 %†—
  • Alibaba · Route 2 · fp8Serving100 %†—
  • Alibaba · Route 3 · fp8Serving96.9 %†—
  • AmbientServing94.8 %†—
  • AtlasCloudServing100 %†—
  • Baidu · Route 1 · fp8Serving98.9 %†—
  • Baidu · Route 2 · fp4Serving100 %†—
  • BaseTen · Route 1 · fp8Serving98.4 %†—
  • BaseTen · Route 2 · fp8Serving99.6 %†—
  • BaseTen · Route 3 · fp8Serving99.5 %†—
  • BaseTen · Route 4 · fp8Serving99.2 %†—
  • CloudflareServing100 %†—
  • CoreWeaveServing99.9 %†—
  • DeepInfra · Route 1 · fp4Serving100 %†—
  • DeepInfra · Route 2 · fp4Serving99.9 %†—
  • DigitalOceanServing99.5 %†—
  • FriendliServing98.4 %†—
  • GMICloudServing100 %†—
  • Hugging Face · deepinfraServing100 %*100 %*
  • InceptronServing99.8 %†—
  • InferHubServing100 %*100 %*
  • Mistral · Route 1 · nvfp4Serving99.7 %†—
  • Mistral · Route 2Serving99.4 %†—
  • MorphServing91.6 %†—
  • NanoGPTServing100 %*100 %*
  • NebiusDown60.4 %†—
  • NovitaServing97.9 %†—
  • ParasailServing96.2 %†—
  • QubaxServing100 %*100 %*
  • RelaceServing100 %†—
  • SandbaseServing100 %*100 %*
  • SiliconFlowServing100 %†—
  • StreamLakeServing99.8 %†—
  • SurplusServing97.4 %*86.5 %*
  • TogetherServing99.2 %†—
  • VeniceServing99.5 %†—
  • Wafer · Route 1 · fp4Serving100 %†—
  • Wafer · Route 2 · fp4Serving100 %†—
  • Z.AIServing100 %†—
GLM 5.1All 16 serving
ProviderNow24 h7 d
  • AlibabaServing95.2 %†—
  • AtlasCloudServing100 %†—
  • BaiduServing100 %†—
  • ChutesServing99.6 %†—
  • FriendliServing99.1 %†—
  • GMICloudServing100 %†—
  • Hugging Face · deepinfraServing100 %*98.3 %*
  • NanoGPTServing100 %*100 %*
  • NovitaServing100 %†—
  • PhalaServing100 %†—
  • QubaxServing97 %*98.8 %*
  • SiliconFlowServing100 %†—
  • StreamLakeServing100 %†—
  • SurplusServing86.2 %*71.5 %*
  • VeniceServing100 %†—
  • Z.AIServing98.1 %†—
GLM 4.7All 11 serving
ProviderNow24 h7 d
  • AiHubMixServing100 %*100 %*
  • DeepInfra · Route 1 · fp4Serving——
  • DeepInfra · Route 2 · fp4Serving99.9 %†—
  • Hugging Face · deepinfraServing100 %*100 %*
  • Mancer 2Serving98.9 %†—
  • NanoGPTServing95.5 %*98.3 %*
  • NovitaServing99.9 %†—
  • QubaxServing100 %*99.4 %*
  • SurplusServing100 %*93.2 %*
  • VeniceServing100 %†—
  • Z.AIServing100 %†—
GLM 4.6All 10 serving
ProviderNow24 h7 d
  • AiHubMixServing100 %*100 %*
  • DeepInfra · Route 1 · fp4Serving100 %†—
  • DeepInfra · Route 2 · fp4Serving100 %†—
  • Hugging Face · deepinfraServing98.4 %*99.4 %*
  • NanoGPTServing100 %*100 %*
  • NovitaServing99.9 %†—
  • QubaxServing100 %*100 %*
  • SurplusServing100 %*94.9 %*
  • VeniceServing100 %†—
  • Z.AIServing100 %†—
DeepSeek V4 Flash3 of 20 degraded
ProviderNow24 h7 d
  • AiHubMixServing100 %*100 %*
  • AtlasCloudServing99.2 %†—
  • AzureServing99.3 %†—
  • BaiduDown86.6 %†—
  • DeepInfraServing100 %†—
  • DigitalOceanServing100 %†—
  • FireworksDown50.5 %†—
  • GMICloudDown96.9 %†—
  • Hugging Face · deepinfraServing100 %*100 %*
  • Mancer 2Serving100 %†—
  • NanoGPTServing100 %*100 %*
  • ParasailServing99.9 %†—
  • QubaxServing100 %*100 %*
  • RelaceServing100 %†—
  • SandbaseServing100 %*100 %*
  • SiliconFlowServing99.4 %†—
  • StreamLakeServing100 %†—
  • SurplusServing100 %*93.6 %*
  • VeniceServing100 %†—
  • WaferServing100 %†—
DeepSeek V4 ProAll 19 serving
ProviderNow24 h7 d
  • AiHubMixServing100 %*100 %*
  • Alibaba · Route 1 · fp8Serving98.2 %†—
  • Alibaba · Route 2 · fp8Serving100 %†—
  • AtlasCloudServing100 %†—
  • AzureServing100 %†—
  • BaiduServing99 %†—
  • DeepInfraServing99.9 %†—
  • DigitalOceanServing100 %†—
  • GMICloudServing99.9 %†—
  • Hugging Face · deepinfraServing100 %*100 %*
  • NanoGPTServing100 %*100 %*
  • NovitaServing100 %†—
  • QubaxServing100 %*100 %*
  • RelaceServing100 %†—
  • SandbaseServing100 %*100 %*
  • SiliconFlowServing100 %†—
  • StreamLakeServing100 %†—
  • SurplusServing98.4 %*93.1 %*
  • VeniceServing99.2 %†—
DeepSeek V4.1 Flash1 of 44 degraded
ProviderNow24 h7 d
  • AiHubMixServing100 %*100 %*
  • AlibabaServing100 %†—
  • AtlasCloudServing100 %†—
  • BaiduServing99.9 %†—
  • BaseTen · Route 1Serving99.5 %†—
  • BaseTen · Route 2Serving99.5 %†—
  • CohereServing99.4 %†—
  • CoreWeaveServing100 %†—
  • CrusoeServing100 %†—
  • DeepInfra · Route 1 · fp8Serving98.4 %†—
  • DeepInfra · Route 2 · fp8Serving98.5 %†—
  • DeepInfra · Route 3 · fp8Down65.5 %†—
  • DigitalOceanServing100 %†—
  • Fireworks · Route 1Serving99.5 %†—
  • Fireworks · Route 2Serving100 %†—
  • Fireworks · Route 3Serving100 %†—
  • GMICloudServing100 %†—
  • GoogleServing98.5 %†—
  • Hugging Face · deepinfraServing96.9 %*98.8 %*
  • InferenceNet · Route 1 · fp8Serving100 %†—
  • InferenceNet · Route 2 · fp8Serving100 %†—
  • InferHubServing95.1 %*95 %
  • kie.aiServing100 %*100 %*
  • MakoraServing100 %†—
  • Mancer 2Serving99.9 %†—
  • ModalServing100 %†—
  • NanoGPTServing100 %*100 %*
  • NovitaServing100 %†—
  • OpenInferenceServing100 %†—
  • Parasail · Route 1 · fp8Serving100 %†—
  • Parasail · Route 2 · fp8Serving100 %†—
  • PhalaServing99.8 %†—
  • PrimeIntellectServing99.4 %†—
  • QubaxServing97 %*98.1 %*
  • RekaServing99.1 %†—
  • RelaceServing100 %†—
  • SandbaseServing100 %*100 %*
  • SiliconFlowServing99.8 %†—
  • StreamLakeServing100 %†—
  • SurplusServing100 %94.5 %
  • TogetherServing100 %†—
  • VeniceServing99.3 %†—
  • Wafer · Route 1 · mxfp4Serving100 %†—
  • Wafer · Route 2 · mxfp4Serving100 %†—
DeepSeek V3.22 of 12 degraded
ProviderNow24 h7 d
  • AtlasCloudServing100 %†—
  • BaiduDown99.6 %†—
  • DeepInfraServing98.6 %†—
  • DigitalOceanDegraded97.3 %†—
  • GMICloudServing100 %†—
  • Hugging Face · deepinfraServing100 %*100 %*
  • NanoGPTServing100 %*100 %*
  • PhalaServing99.8 %†—
  • QubaxServing100 %*98.3 %*
  • SandbaseServing100 %*99.4 %*
  • SiliconFlowServing99.1 %†—
  • SurplusServing96.6 %*89.8 %*
deepseek-ai/DeepSeek-R1All 2 serving
ProviderNow24 h7 d
  • Hugging Face · novitaServing—98.9 %*
  • QubaxServing0 %*0 %*
Qwen3.8 FlashAll 7 serving
ProviderNow24 h7 d
  • AiHubMixServing100 %*93.2 %*
  • Alibaba · Route 1Serving100 %†—
  • Alibaba · Route 2Serving100 %†—
  • InferHubServing95.5 %*98.1 %*
  • NanoGPTServing100 %*99.4 %*
  • QubaxServing100 %*57.4 %
  • SurplusServing98.3 %*87.9 %
Qwen 3 Coder 480B1 of 5 degraded
ProviderNow24 h7 d
  • DeepInfraDown92.5 %†—
  • NanoGPTServing92.3 %*94.8 %*
  • QubaxServing100 %*100 %*
  • SurplusServing—0 %*
  • VeniceServing88.5 %†—
Qwen3 Coder NextAll 4 serving
ProviderNow24 h7 d
  • NanoGPTServing100 %*100 %*
  • ParasailServing99.9 %†—
  • QubaxServing100 %*99.4 %*
  • SurplusServing——
Qwen3 235B A22B Instruct 2507All 8 serving
ProviderNow24 h7 d
  • DeepInfraServing86.3 %†—
  • GMICloudServing0 %†—
  • Hugging Face · deepinfraServing96.8 %*98.2 %*
  • NanoGPTServing100 %*100 %*
  • ParasailServing100 %†—
  • QubaxServing100 %*100 %*
  • SurplusServing100 %*93.8 %*
  • VeniceServing92.5 %†—
Qwen3.6 27BAll 12 serving
ProviderNow24 h7 d
  • AlibabaServing100 %†—
  • ChutesServing91.9 %†—
  • DeepInfraServing99.7 %†—
  • Hugging Face · deepinfraServing98.5 %*99.4 %*
  • NanoGPTServing100 %*100 %*
  • PhalaServing82.1 %†—
  • QubaxServing100 %*99.2 %*
  • RekaServing99.2 %†—
  • SandbaseServing—100 %*
  • SiliconFlowServing94 %†—
  • SurplusServing100 %*93.8 %*
  • VeniceServing97.6 %†—
Qwen/Qwen3-32B1 of 8 degraded
ProviderNow24 h7 d
  • DeepInfraServing99.7 %†—
  • Hugging Face · deepinfraServing100 %*98.8 %*
  • NanoGPTServing95.5 %*97.1 %*
  • QubaxServing100 %*100 %*
  • SandbaseServing100 %*99.4 %*
  • SiliconFlowServing91.3 %†—
  • SurplusServing100 %*92.5 %*
  • TenstorrentDegraded93.4 %†—
Llama 3.3 70B InstructAll 15 serving
ProviderNow24 h7 d
  • AiHubMixServing100 %*100 %*
  • AkashMLServing100 %†—
  • CloudflareServing100 %†—
  • CoreWeaveServing100 %†—
  • DeepInfraServing99.5 %†—
  • Google · Route 1Serving——
  • Google · Route 2Serving——
  • GroqServing100 %†—
  • Hugging Face · novitaServing100 %*100 %*
  • NanoGPTServing100 %*98.8 %*
  • ParasailServing100 %†—
  • QubaxServing100 %*99.4 %*
  • SambaNovaServing99.6 %†—
  • SandbaseServing100 %*100 %*
  • SurplusServing—0 %*
Llama 4 MaverickAll 5 serving
ProviderNow24 h7 d
  • AiHubMixServing100 %*100 %*
  • DigitalOceanServing100 %†—
  • GoogleServing——
  • NanoGPTServing100 %*100 %*
  • ParasailServing100 %†—
Llama 4 ScoutAll 6 serving
ProviderNow24 h7 d
  • AiHubMixServing100 %*100 %*
  • DeepInfra · Route 1 · fp8Serving100 %†—
  • DeepInfra · Route 2 · fp8Serving100 %†—
  • DeepInfra · Route 3 · fp8Serving99.3 %†—
  • GoogleServing——
  • NanoGPTServing100 %*100 %*
Gemma 4 26B A4B ItAll 20 serving
ProviderNow24 h7 d
  • AiHubMixServing100 %*100 %*
  • AkashMLServing100 %†—
  • CloudflareServing100 %†—
  • CoreWeaveServing97.8 %†—
  • DarkbloomServing100 %†—
  • DeepInfraServing100 %†—
  • DekaLLMServing100 %†—
  • GoogleServing97.1 %†—
  • Hugging Face · deepinfraServing100 %*100 %*
  • Io NetServing99.8 %†—
  • MakoraServing100 %†—
  • NanoGPTServing98.5 %*98.8 %*
  • NextBitServing100 %†—
  • NovitaServing93.6 %†—
  • ParasailServing100 %†—
  • PrimeIntellectServing100 %†—
  • QubaxServing100 %*100 %*
  • SiliconFlowServing99.6 %†—
  • SurplusServing100 %*94.9 %*
  • VeniceServing100 %†—
google/gemma-3-27b-itAll 8 serving
ProviderNow24 h7 d
  • DeepInfraServing100 %†—
  • Hugging Face · deepinfraServing100 %*100 %*
  • NanoGPTServing100 %*100 %*
  • NebiusServing97.3 %†—
  • NextBitServing96.3 %†—
  • ParasailServing100 %†—
  • QubaxServing100 %*100 %*
  • SurplusServing100 %*92.6 %*
gpt-oss-120b1 of 32 degraded
ProviderNow24 h7 d
  • AiHubMixServing100 %*100 %*
  • AkashMLServing100 %†—
  • Amazon Bedrock · Route 1Serving98.2 %†—
  • Amazon Bedrock · Route 2Serving97 %†—
  • BaseTen · Route 1 · fp4Serving100 %†—
  • BaseTen · Route 2 · fp4Serving100 %†—
  • CerebrasServing100 %†—
  • CoreWeaveServing100 %†—
  • CrusoeServing99.8 %†—
  • DeepInfra · Route 1 · bf16Serving100 %†—
  • DeepInfra · Route 2 · fp8Serving75.2 %†—
  • DekaLLMServing100 %†—
  • DigitalOceanServing100 %†—
  • GoogleDown4.4 %†—
  • GroqServing92.9 %†—
  • Hugging Face · deepinfraServing95.1 %*97.6 %*
  • Mancer 2Serving97.9 %†—
  • NanoGPTServing94 %*97.1 %*
  • Near AIServing100 %†—
  • NebiusServing100 %†—
  • Parasail · Route 1 · fp4Serving100 %†—
  • Parasail · Route 2 · fp4Serving100 %†—
  • PhalaServing99.5 %†—
  • PrimeIntellectServing100 %†—
  • QubaxServing100 %*100 %*
  • Sail ResearchServing100 %†—
  • SambaNovaServing99.2 %†—
  • SandbaseServing100 %*100 %*
  • SiliconFlowServing77.7 %†—
  • SurplusServing100 %*94.9 %*
  • TogetherServing100 %†—
  • VeniceServing100 %†—
MiniMax M3All 17 serving
ProviderNow24 h7 d
  • AiHubMixServing——
  • CoreWeaveServing100 %†—
  • DeepInfra · Route 1 · fp8Serving100 %†—
  • DeepInfra · Route 2 · fp8Serving100 %†—
  • GMICloudServing100 %†—
  • Hugging Face · deepinfraServing——
  • MinimaxServing100 %†—
  • ModelRunServing100 %†—
  • NanoGPTServing——
  • NovitaServing100 %†—
  • Parasail · Route 1 · fp8Serving99.7 %†—
  • Parasail · Route 2 · fp8Serving100 %†—
  • QubaxServing——
  • SandbaseServing——
  • StreamLakeServing100 %†—
  • SurplusServing——
  • VeniceServing98.8 %†—
MiniMax M2.7All 8 serving
ProviderNow24 h7 d
  • AiHubMixServing——
  • GMICloudServing99.9 %†—
  • GroqServing99.6 %†—
  • Hugging Face · deepinfraServing——
  • NanoGPTServing——
  • QubaxServing——
  • SandbaseServing——
  • SurplusServing——
Mistral Small 3.2 24B InstructAll 7 serving
ProviderNow24 h7 d
  • DeepInfraServing100 %†—
  • MistralServing99.9 %†—
  • NanoGPTServing——
  • ParasailServing98.8 %†—
  • QubaxServing——
  • SurplusServing——
  • VeniceServing98.6 %†—
Xiaomi Mimo V2.6 Pro9 of 10 serving
ProviderNow24 h7 d
  • AiHubMixServing——
  • DeepInfraServing100 %†—
  • GMICloudServing88.9 %†—
  • NanoGPTServing——
  • PrimaLabsServing99.4 %†—
  • QubaxServing——
  • SandbaseServing——
  • StreamLakeServing94.4 %†—
  • SurplusNot taking requests——
  • XiaomiServing100 %†—
Nvidia Nemotron 3 Super 120B A12BAll 7 serving
ProviderNow24 h7 d
  • AiHubMixServing——
  • CrusoeServing99.5 %†—
  • DeepInfraServing100 %†—
  • DekaLLMServing99.2 %†—
  • NanoGPTServing——
  • QubaxServing——
  • SurplusServing——
Claude Sonnet 5All 14 serving
ProviderNow24 h7 d
  • AiHubMixServing100 %*100 %*
  • Amazon Bedrock · Route 1Serving96.6 %†—
  • Amazon Bedrock · Route 2Serving100 %†—
  • Amazon Bedrock · Route 3Serving100 %†—
  • AnthropicServing100 %†—
  • Azure · Route 1Serving100 %†—
  • Azure · Route 2Serving——
  • Claude Platform on AWSServing100 %†—
  • Google · Route 1Serving99.5 %†—
  • Google · Route 2Serving99.4 %†—
  • Google · Route 3Serving100 %†—
  • NanoGPTServing100 %*100 %*
  • QubaxServing97.3 %*98.6 %*
  • SurplusServing100 %*95.5 %*
Claude Opus 5.5All 15 serving
ProviderNow24 h7 d
  • AiHubMixServing100 %*100 %*
  • Amazon Bedrock · Route 1Serving100 %†—
  • Amazon Bedrock · Route 2Serving100 %†—
  • Amazon Bedrock · Route 3Serving100 %†—
  • Anthropic · Route 1Serving100 %†—
  • Anthropic · Route 2Serving100 %†—
  • Azure · Route 1Serving99.1 %†—
  • Azure · Route 2Serving100 %†—
  • Claude Platform on AWSServing100 %†—
  • Google · Route 1Serving100 %†—
  • Google · Route 2Serving100 %†—
  • Google · Route 3Serving100 %†—
  • NanoGPTServing100 %*100 %*
  • QubaxServing100 %*98.3 %*
  • SurplusServing100 %*98.3 %*
GPT 5.5All 12 serving
ProviderNow24 h7 d
  • AiHubMixServing100 %*100 %*
  • Amazon BedrockServing——
  • Azure · Route 1Serving99.6 %†—
  • Azure · Route 2Serving——
  • Azure · Route 3Serving——
  • NanoGPTServing100 %*100 %*
  • OpenAI · Route 1Serving100 %†—
  • OpenAI · Route 2Serving99.8 %†—
  • OpenAI · Route 3Serving——
  • QubaxServing95.6 %*98.3 %*
  • SandbaseServing100 %*100 %*
  • SurplusServing100 %*91 %*
Gemini 3.5 FlashAll 11 serving
ProviderNow24 h7 d
  • Google · Route 1Serving100 %†—
  • Google · Route 2Serving100 %†—
  • Google · Route 3Serving——
  • Google · Route 4Serving100 %†—
  • Google AI Studio · Route 1Serving100 %†—
  • Google AI Studio · Route 2Serving100 %†—
  • Google AI Studio · Route 3Serving100 %†—
  • NanoGPTServing77.2 %*80 %*
  • QubaxServing100 %*99.4 %*
  • SandbaseServing100 %*100 %*
  • SurplusServing100 %*94.3 %*
Grok 4.71 of 12 degraded
ProviderNow24 h7 d
  • AiHubMixServing100 %*100 %*
  • Amazon BedrockServing97 %†—
  • kie.aiServing94.1 %*79.3 %*
  • NanoGPTServing100 %*100 %*
  • QubaxServing100 %*98.3 %*
  • SandbaseServing—100 %*
  • SurplusServing100 %*88.1 %*
  • xAI · Route 1Serving99 %†—
  • xAI · Route 2Degraded96.4 %†—
  • xAI · Route 3Serving100 %†—
  • xAI · Route 4Serving——
  • xAI · Route 5Serving100 %†—

Uptime is the share of minutes in which a route answered at least 80 % of the requests sent to it; a minute counts once it has 10 requests or more. * Where a route had too few requests a minute, its figure is a success rate: the share of its requests answered. † A published figure for this host, not one measured on our requests. — No figure for that window yet.

Degraded: down, or under 90 % in the last 30 minutes. Not taking requests: the route has no offer at its price right now; it is not counted as degraded, and requests go to the other routes.