PRICEDeepSeek V4 Flash via OpenRouter: output price ↓ 33%1mPRICEDeepSeek V4 Flash via OpenRouter: input price ↓ 33%1mPRICE~moonshotai/kimi-latest input_per_mtok decreased from 1.39 to 1.342mPRICE~deepseek/deepseek-v4-flash-latest cached_input_per_mtok decreased from 0.0086 to 0.00382mPRICE~deepseek/deepseek-v4-flash-latest input_per_mtok decreased from 0.0086 to 0.00382mPRICEdeepseek/deepseek-v4-flash-0731 cached_input_per_mtok decreased from 0.012 to 0.00512mPRICEdeepseek/deepseek-v4-flash-0731 input_per_mtok decreased from 0.012 to 0.00512mPRICEdeepseek/deepseek-v4-flash cached_input_per_mtok decreased from 0.0084 to 0.00562mPRICEdeepseek/deepseek-v4-flash output_per_mtok decreased from 0.084 to 0.0562mPRICEdeepseek/deepseek-v4-flash input_per_mtok decreased from 0.042 to 0.0282mPRICE~deepseek/deepseek-v4-flash-latest cached_input_per_mtok decreased from 0.013 to 0.008633mPRICE~deepseek/deepseek-v4-flash-latest input_per_mtok decreased from 0.013 to 0.008633mCAPABILITYxiaomi/mimo-v2.6-flash now declares support for top_logprobs33mCAPABILITYxiaomi/mimo-v2.6-flash now declares support for logprobs33mPRICEdeepseek/deepseek-v4-flash-0731 cached_input_per_mtok decreased from 0.017 to 0.01233mPRICEdeepseek/deepseek-v4-flash-0731 input_per_mtok decreased from 0.017 to 0.01233mPRICE~deepseek/deepseek-v4-flash-latest cached_input_per_mtok increased from 0.0058 to 0.0131hPRICE~deepseek/deepseek-v4-flash-latest input_per_mtok increased from 0.0058 to 0.0131hCAPABILITYopenai/gpt-5.4-image-2 no longer declares support for stop1hCAPABILITYopenai/gpt-5.4-image-2 no longer declares support for presence_penalty1hCAPABILITYopenai/gpt-5.4-image-2 no longer declares support for logit_bias1hCAPABILITYopenai/gpt-5.4-image-2 no longer declares support for frequency_penalty1hCAPABILITYopenai/gpt-5-image-mini no longer declares support for stop1hCAPABILITYopenai/gpt-5-image-mini no longer declares support for presence_penalty1hCAPABILITYopenai/gpt-5-image-mini no longer declares support for logit_bias1hCAPABILITYopenai/gpt-5-image-mini no longer declares support for frequency_penalty1hCAPABILITYopenai/gpt-5-image no longer declares support for stop1hCAPABILITYopenai/gpt-5-image no longer declares support for presence_penalty1hCAPABILITYopenai/gpt-5-image no longer declares support for logit_bias1hCAPABILITYopenai/gpt-5-image no longer declares support for frequency_penalty1hPRICEdeepseek/deepseek-v4-flash-0731 cached_input_per_mtok increased from 0.0077 to 0.0171hPRICEdeepseek/deepseek-v4-flash-0731 input_per_mtok increased from 0.0077 to 0.0171hPRICE~deepseek/deepseek-v4-flash-latest cached_input_per_mtok decreased from 0.0086 to 0.00582hPRICE~deepseek/deepseek-v4-flash-latest input_per_mtok decreased from 0.0086 to 0.00582hPRICEdeepseek/deepseek-v4-flash-0731 cached_input_per_mtok decreased from 0.012 to 0.00772hPRICEdeepseek/deepseek-v4-flash-0731 input_per_mtok decreased from 0.012 to 0.00772hPRICE~deepseek/deepseek-v4-flash-latest cached_input_per_mtok decreased from 0.013 to 0.00862hPRICE~deepseek/deepseek-v4-flash-latest input_per_mtok decreased from 0.013 to 0.00862hPRICEdeepseek/deepseek-v4-flash-0731 cached_input_per_mtok decreased from 0.017 to 0.0122hPRICEdeepseek/deepseek-v4-flash-0731 input_per_mtok decreased from 0.017 to 0.0122h

Cortecs vs Groq

2 shared models · AT vs US · independent benchmarks

Registry facts

FactCortecsGroq
Typeai gatewayinference host
HeadquartersATUS
Catalogued endpoints11
EU residencyYesNo

GPT-OSS 120B: Cortecs vs Groq

● MEASURED Cortecs detail → · ● MEASURED Groq detail →

MetricCortecsGroq
TTFT p50240 ms264 ms
TTFT p95501 ms402 ms
Throughput358 tok/s486 tok/s
Reliability100%100%
Tool calling100%100%
Structured output100%100%
$ output / M tokens$0.46$0.60

GPT-OSS 20B: Cortecs vs Groq

● MEASURED Cortecs detail → · ● MEASURED Groq detail →

MetricCortecsGroq
TTFT p50175 ms258 ms
TTFT p95588 ms367 ms
Throughput67 tok/s929 tok/s
Reliability99.5%100%
Tool calling100%100%
Structured output100%100%
$ output / M tokens$0.17$0.30

Every figure above is an independent measurement from the canonical vantage; unmeasured fields show — until measured. How we benchmark →