Pricing Update

OpenAI

• New model added: gpt-6.1-sol.

  • Standard: same input/output pricing as gpt-6-sol, but cached input is cheaper:
    • short cached input: $0.10 vs gpt-6-sol $0.20
    • long cached input: $0.20 vs gpt-6-sol $0.40
  • Batch:
    • input/output same as gpt-6-sol
    • short cached input: $0.05 vs $0.10
    • long cached input: $0.10 vs $0.20
  • Flex:
    • input/output same as gpt-6-sol
    • short cached input: $0.05 vs $0.10
    • long cached input: $0.10 vs $0.20
  • Fast:
    • input/output same as gpt-6-sol
    • short cached input: $0.20 vs $0.40
    • long cached input: $0.40 vs $0.80

• New pricing tier added: Ultrafast.

  • Only listed model: gpt-6-astra
  • Prices:
    • short input $60.00
    • short cached input $6.00
    • short cache writes $75.00
    • short output $300.00
    • long input $120.00
    • long cached input $12.00
    • long cache writes $150.00
    • long output $450.00

• Billing/surcharge wording changed for uplifts:

  • NEW says:
    • regional processing endpoints are charged a 10% uplift for models released on or after March 5, 2026
    • FedRAMP endpoints are also charged a 10% uplift
  • OLD said FedRAMP uplift applied over corresponding standard model rates, and regional processing uplift applied only to models eligible for data residency.
  • Substantively, the document now states the two 10% uplifts together in the flagship section, while removing the earlier Bedrock-specific pricing equivalence note from that section.

• Cloud platform billing note added as a separate section:

  • OpenAI models on Amazon Bedrock and Microsoft Azure are billed through those services.
  • OLD only mentioned Bedrock billing through AWS and said Bedrock commercial-region pricing matches direct OpenAI pricing. That pricing-equivalence statement is removed; Azure billing is newly mentioned.

• Model/tier naming text changed:

  • Fast mode is now labeled Fast in section headers.
  • This is not a price change, but the billing note still says priority processing was renamed Fast mode.

• Daybreak alias billing note removed:

  • OLD said gpt-daybreak-blue-latest mapped to gpt-5.6-sol and gpt-daybreak-red-latest mapped to gpt-5.6-cyber, with pricing following the underlying model.
  • This alias/pricing rule is absent in NEW.

• Data residency availability notes removed from flagship pricing sections:

  • OLD specified:
    • for GPT-6 Sol and Luna, EU data residency is available only with Standard processing
    • for Fast mode, GPT-6 Astra, Sol, and Luna EU data residency is available only with Standard processing
  • These billing-related availability constraints are not stated in NEW.