Revenir au contenu principal

Proprietary Foundation Model Serving

Servez des modèles de fondation ouverts haut de gamme pour vos charges de travail d'inférence en batch et en temps réel. Cela vous permet de créer rapidement et facilement des applications qui exploitent des modèles d’AI générative propriétaires de haute qualité, proposés par divers fournisseurs, directement sur la plateforme Databricks, sans démarches supplémentaires ni contacts séparés avec d’autres fournisseurs.

Loading...

Proprietary Foundation Model Serving DBU rates

Model Pay-Per-Token
 DBU / 1M INPUT tokens
(Global)
DBU / 1M OUTPUT tokens
(Global)
OpenAI
GPT 5Global17.857142.857
In-geo19.643157.143
GPT 5 MiniGlobal3.57128.571
In-geo3.92931.429
GPT 5 NanoGlobal0.7145.714
In-geo0.7866.286
Anthropic
Claude Opus 4.1Global214.2861,071.43
Claude Sonnet 4.5Global42.857214.286
In-geo47.143235.715
Claude Sonnet 4Global42.857214.286
Claude Sonnet 3.7Global42.857214.286

Proprietary Foundation Model Serving DBU rates

ModelEndpoint typeContext Length

Pay Per Token

Batch Inference
InputOutputCache writesCache reads 
DBU / 1M TokensDBU / 1M TokensDBU / 1M TokensDBU / 1M TokensDBU / hour
OpenAI
GPT 5.2GlobalAll Lengths25.000200.00025.0002.500184.286
In-geo27.500220.00027.5002.750202.714
GPT 5.1GlobalAll Lengths17.857142.85717.8571.786131.429
In-geo19.643157.14319.6431.965144.571
GPT 5.1 Codex MaxGlobalAll Lengths17.857142.85717.8571.786131.429
In-geo19.643157.14319.6431.965144.571
GPT 5GlobalAll Lengths17.857142.85717.8571.786131.429
In-geo19.643157.14319.6431.965144.571
GPT 5 miniGlobalAll Lengths3.57128.5713.5710.35771.429
In-geo3.92931.4293.9290.39378.571
GPT 5.1 Codex MiniGlobalAll Lengths3.57128.5713.5710.35771.429
In-geo3.92931.4293.9290.39378.571
GPT 5 nanoGlobalAll Lengths0.7145.7140.7140.07153.571
In-geo0.7866.2860.7860.07858.929

Proprietary Foundation Model Serving DBU rates

ModelEndpoint typeContext Length

Pay Per Token

Batch Inference
InputOutputCache writesCache reads 
DBU / 1M TokensDBU / 1M TokensDBU / 1M TokensDBU / 1M TokensDBU / hour
Anthropic
Claude Opus 4.6GlobalShort Context71.429357.14389.2867.143178.571
In-geo78.571392.85798.2147.857196.429
GlobalLong Context
(>200k tokens)
142.858535.715178.57214.286178.571
In-geo157.142589.286196.42815.714196.429
Claude Opus 4.5GlobalShort Context71.429357.14389.2867.143178.571
In-geo78.571392.85798.2147.857196.429
Claude Opus 4 / 4.1Global/In-geoAll Lengths214.2861,071.429267.85721.429514.286
Claude Sonnet 4.5 / 4.6GlobalShort Context42.857214.28653.5714.286214.286
In-geo47.143235.71558.9284.715235.715
GlobalLong Context
(>200k tokens)
85.714321.429107.1438.571214.286
In-geo94.285353.572117.8579.428235.715
Claude Sonnet 3.7 / 4 / 4.1

Claude 3.7 Sonnet will be deprecated on April 12, 2026
Global/In-geoShort Context42.857214.28653.5714.286214.286
Long Context
(>200k tokens)
85.714321.429107.1438.571214.286
Claude Haiku 4.5GlobalAll Lengths14.28671.42917.8571.429114.286
In-geo15.71578.57219.6431.572125.714

Proprietary Foundation Model Serving DBU rates

ModelEndpoint typeContext Length

Pay Per Token

Batch Inference
InputOutputCache writesCache reads 
DBU / 1M TokensDBU / 1M TokensDBU / 1M TokensDBU / 1M TokensDBU / hour
Google
Gemini 3.0/3.1 ProGlobal/In-geoShort Context35.714214.28635.7143.571230.357
Long Context
(>200k tokens)
71.429321.42971.4297.143230.357
Gemini 3.0 FlashGlobal/In-geoShort Context8.92953.5718.9290.893125.000
Long Context
(>200k tokens)
8.92953.5718.9290.893125.000
Gemini 2.5 ProGlobal/In-geoShort Context17.857142.857n/an/a164.286
Long Context
(>200k tokens)
35.714214.286n/an/a164.286
Gemini 2.5 FlashGlobal/In-geoShort Context4.28635.714n/an/a107.143
Long Context
(>200k tokens)
4.28635.714n/an/a107.143

Payez à l'utilisation avec un essai gratuit de 14 jours. Ou contactez-nous pour connaître les remises sur engagements de dépenses et nous détailler vos besoins spécifiques.

FAQ sur le Service du Modèle de Fondation Partenaire