PRICE~moonshotai/kimi-latest cached_input_per_mtok increased from 0.29 to 0.8028mPRICE~moonshotai/kimi-latest output_per_mtok increased from 11.36 to 13.0028mPRICE~moonshotai/kimi-latest input_per_mtok increased from 0.75 to 0.9928mPERFORMANCEGLM-5.2 → Scaleway: TTFT ↑ 20%58mPERFORMANCEDeepSeek V4 Flash → Scaleway: TTFT ↑ 35%58mPERFORMANCEMistral Medium 3.5 → Scaleway: TTFT ↓ 21%58mPERFORMANCEDeepSeek V4 Flash → Scaleway: throughput ↓ 20%58mPERFORMANCEGPT-OSS 120B → Scaleway: throughput ↑ 31%58mPERFORMANCEGPT-OSS 120B → Scaleway: TTFT ↓ 24%58mPERFORMANCEGPT-OSS 120B → Together AI: throughput ↑ 20%58mPERFORMANCEKimi K3 via OpenRouter: TTFT ↓ 49%58mPERFORMANCEGLM-5.2 via OpenRouter: reliability recovered 96.8% → 100.0%58mPERFORMANCEKimi K3 via Cortecs: throughput ↑ 184%58mPERFORMANCEKimi K3 via Cortecs: TTFT ↓ 42%58mPERFORMANCEGLM-5.2 via OpenRouter: throughput ↓ 49%58mPERFORMANCEGLM-5.2 via OpenRouter: TTFT ↑ 121%58mPERFORMANCEMiniMax M3 via OpenRouter: reliability recovered 96.8% → 100.0%58mPERFORMANCEDeepSeek V4 Pro via Cortecs: throughput ↑ 27%58mPERFORMANCEGPT-OSS 120B via OpenRouter: throughput ↓ 22%58mPERFORMANCEGLM-4.7 via OpenRouter: TTFT ↓ 78%58mPERFORMANCEGPT-OSS 20B → Groq: TTFT ↑ 24%58mPERFORMANCEKimi K3 → Together AI: TTFT ↓ 34%58mPERFORMANCEDeepSeek V4 Flash via OpenRouter: throughput ↓ 17%58mPERFORMANCELlama 3.3 70B via Cortecs: TTFT ↓ 51%58mPERFORMANCEQwen3 235B via OpenRouter: throughput ↑ 37%58mPERFORMANCEQwen3 235B via OpenRouter: TTFT ↓ 29%58mPERFORMANCEDeepSeek V4 Flash → Together AI: throughput ↑ 19%58mPERFORMANCEGLM-5.3 via OpenRouter: throughput ↓ 17%58mPERFORMANCEGPT-OSS 20B via Cortecs: throughput ↑ 28%58mPERFORMANCELlama 3.3 70B → Together AI: TTFT ↓ 33%58mPRICE~moonshotai/kimi-latest cached_input_per_mtok decreased from 0.80 to 0.2959mPRICE~moonshotai/kimi-latest output_per_mtok decreased from 13.00 to 11.3659mPRICE~moonshotai/kimi-latest input_per_mtok decreased from 0.99 to 0.7559mPRICE~deepseek/deepseek-v4-flash-latest cached_input_per_mtok decreased from 0.0077 to 0.001159mPRICE~deepseek/deepseek-v4-flash-latest output_per_mtok decreased from 1.28 to 1.0459mPRICE~deepseek/deepseek-v4-flash-latest input_per_mtok decreased from 0.0077 to 0.003859mPRICEz-ai/glm-5.1 cached_input_per_mtok increased from 0.18 to 0.2659mPRICEz-ai/glm-5.1 output_per_mtok increased from 3.03 to 4.4059mPRICEz-ai/glm-5.1 input_per_mtok increased from 0.96 to 1.4059mPRICEtencent/hy4-preview cached_input_per_mtok increased from 0.038 to 0.04259m
Not yet benchmarked

OpenInference is in the registry, awaiting benchmark coverage. View the live ranking →

OpenInference in the registry

  • HeadquartersUS
  • API protocols—
  • Catalogued endpoints0

All providers · How we benchmark