PRICE~z-ai/glm-flash-latest cached_input_per_mtok decreased from 0.013 to 0.0118mPRICE~z-ai/glm-flash-latest output_per_mtok increased from 0.81 to 0.908mPRICE~z-ai/glm-flash-latest input_per_mtok increased from 0.014 to 0.0298mPRICE~moonshotai/kimi-latest cached_input_per_mtok increased from 0.31 to 0.338mPRICE~moonshotai/kimi-latest input_per_mtok increased from 0.32 to 0.348mPRICE~deepseek/deepseek-v4-flash-latest input_per_mtok decreased from 0.016 to 0.0128mPRICE~deepseek/deepseek-flash-latest cached_input_per_mtok decreased from 0.01 to 0.00548mPRICE~deepseek/deepseek-flash-latest output_per_mtok decreased from 0.80 to 0.728mPRICE~deepseek/deepseek-flash-latest input_per_mtok decreased from 0.017 to 0.00938mPRICEdeepseek/deepseek-v4-pro cached_input_per_mtok decreased from 0.03 to 0.0298mPRICEdeepseek/deepseek-v4-pro output_per_mtok decreased from 0.71 to 0.708mPRICEdeepseek/deepseek-v4-pro input_per_mtok decreased from 0.35 to 0.358mPRICEdeepseek/deepseek-v4-flash-0731 cached_input_per_mtok decreased from 0.018 to 0.0138mPRICEdeepseek/deepseek-v4-flash-0731 input_per_mtok decreased from 0.018 to 0.0138mPRICE~z-ai/glm-flash-latest cached_input_per_mtok increased from 0.011 to 0.01339mPRICE~z-ai/glm-flash-latest output_per_mtok decreased from 0.90 to 0.8139mPRICE~z-ai/glm-flash-latest input_per_mtok decreased from 0.016 to 0.01439mPRICE~moonshotai/kimi-latest cached_input_per_mtok increased from 0.27 to 0.3139mPRICE~moonshotai/kimi-latest input_per_mtok increased from 0.28 to 0.3239mPRICE~deepseek/deepseek-v4-flash-latest cached_input_per_mtok increased from 0.0051 to 0.01139mPRICE~deepseek/deepseek-v4-flash-latest output_per_mtok decreased from 0.64 to 0.5839mPRICE~deepseek/deepseek-v4-flash-latest input_per_mtok increased from 0.0051 to 0.01639mPRICE~deepseek/deepseek-flash-latest cached_input_per_mtok increased from 0.0054 to 0.0139mPRICE~deepseek/deepseek-flash-latest output_per_mtok increased from 0.72 to 0.8039mPRICE~deepseek/deepseek-flash-latest input_per_mtok increased from 0.011 to 0.01739mPRICEz-ai/glm-5.3-flash cached_input_per_mtok increased from 0.011 to 0.0339mPRICEz-ai/glm-5.3-flash output_per_mtok decreased from 0.90 to 0.5039mPRICEz-ai/glm-5.3-flash input_per_mtok increased from 0.016 to 0.1539mPRICEmoonshotai/kimi-k3 cached_input_per_mtok increased from 0.18 to 0.2839mPRICEmoonshotai/kimi-k3 output_per_mtok increased from 12.00 to 15.0039mPRICEmoonshotai/kimi-k3 input_per_mtok increased from 0.29 to 0.4139mPRICEdeepseek/deepseek-v4-pro cached_input_per_mtok decreased from 0.03 to 0.0339mPRICEdeepseek/deepseek-v4-pro output_per_mtok decreased from 0.72 to 0.7139mPRICEdeepseek/deepseek-v4-pro input_per_mtok decreased from 0.36 to 0.3539mPRICEdeepseek/deepseek-v4-flash-0731 cached_input_per_mtok increased from 0.0051 to 0.01839mPRICEdeepseek/deepseek-v4-flash-0731 input_per_mtok increased from 0.0051 to 0.01839mNEW MODELmoondream/moondream3.1-9b-a2b appeared in the catalogue39mNEW MODELmoondream/moondream3-preview appeared in the catalogue39mPRICE~z-ai/glm-flash-latest input_per_mtok decreased from 0.02 to 0.0161hPRICE~moonshotai/kimi-latest cached_input_per_mtok increased from 0.18 to 0.271h

Same model. Different infrastructure. Different result.

Global provider ranking

92 providers · 1035 models · 26 endpoints
#1 OverallCerebras99.0FastestCerebras0.43 sBest valueGroqEurope #1Mistral98.9Best toolsCerebras100.0%Best structuredCerebras100.0%
#ProviderΔ30d
01CerebrasBEST OVERALL● MEASUREDUS · Serverless · 1 model · 1 region100.0%100.0%100.0%1896250.4s1687P86↓ 1.8
02MistralBEST EUROPEANEU RESIDENT● MEASUREDFR · Serverless · 2 models · 1 region100.0%100.0%91.7%35419394.4s110P29—
03GroqBEST VALUE● MEASUREDUS · Serverless · 2 models · 1 region100.0%100.0%100.0%2644430.9s697P100↓ 0.3
04OpenAISOLE SERVER● MEASUREDUS · Serverless · 2 models · 1 region100.0%100.0%100.0%110821745.3s97P14—
05ScalewayEU RESIDENTEU OWNED● MEASUREDFR · Serverless · 6 models · 1 region100.0%99.7%100.0%1889555.3s91P43↑ 4.1
06OpenRouter● MEASUREDUS · Routed · 13 models · 2 regions99.4%——741———P71—
07Together● MEASUREDUS · Serverless · 6 models · 1 region97.4%97.6%96.8%54222545.9s111P57↓ 2.6
7 of 18 providers ranked · · Methodology →
01Cerebras99.0
Best overallUS · Serverless · 1 model · 1 region
TTFT189 ms
Rel.100.0%
Tok/s1687
$ Out$0.75
Tools100.0%
Δ30d-1.8
View dossier
02Mistral98.9
Best EuropeanFR · Serverless · 2 models · 1 region
TTFT354 ms
Rel.100.0%
Tok/s110
$ Out$4.50
Tools100.0%
Δ30d—
View dossier
03Groq98.3
Best valueUS · Serverless · 2 models · 1 region
TTFT264 ms
Rel.100.0%
Tok/s697
$ Out$0.45
Tools100.0%
Δ30d-0.3
View dossier
04OpenAI92.9
US · Serverless · 2 models · 1 region
TTFT1108 ms
Rel.100.0%
Tok/s97
$ Out$15.60
Tools100.0%
Δ30d—
View dossier
05Scaleway92.6
FR · Serverless · 6 models · 1 region
TTFT188 ms
Rel.100.0%
Tok/s91
$ Out$1.81
Tools99.7%
Δ30d+4.1
View dossier
06OpenRouter86.7
US · Routed · 13 models · 2 regions
TTFT741 ms
Rel.99.4%
Tok/s—
$ Out$1.20
Tools—
Δ30d—
View dossier
07Together81.1
US · Serverless · 6 models · 1 region
TTFT542 ms
Rel.97.4%
Tok/s111
$ Out$1.12
Tools97.6%
Δ30d-2.6
View dossier
7 of 18 providers ranked · · Methodology →
Benchmark Partners · Become a Founding Partner →

Provider discovery

All 92 providers →
Recently benchmarked
European providersEuropean infrastructure →
Registry

92 providers · 1035 observed models · 26 catalogued endpoints, live from continuous discovery.

How we benchmark →