All models
Text & reasoningLive
OpenAIDaybreak Blue
OpenAI's approval-gated access to its flagship models for defensive security work, with fewer cybersecurity refusals.
Plain-English overview
What Daybreak Blue actually is
Daybreak Blue is an access tier in OpenAI's Daybreak program, not a separately trained model. The API alias currently points to GPT-5.6 Sol, so it shares that model's 1.05M-token context, 128K-token output limit and February 16, 2026 knowledge cutoff. Approved users can also apply Daybreak Blue to GPT-6 Sol and GPT-6 Luna.
It relaxes system-level cyber safeguards for verified defenders, but the model can still refuse highly dual-use requests. Exploit development and penetration testing require the separate Daybreak Red approval. Access is tied to an approved identity, workspace or API project and stays off until an administrator enables it.
Good fit for
- Vulnerability triage
- Malware analysis in a sandbox
- Incident response and detection rules
Category comparison
The facts that matter for text models
These are provider-published specifications, not Cody benchmark scores. Follow the linked sources for current limits and endpoint-specific exceptions.
- Context window
- 1.05M tokensMaximum combined prompt and working context documented by the provider.
- Maximum output
- 128K tokensProvider-published response limit, where available.
- Knowledge cutoff
- February 16, 2026Latest reliable knowledge date explicitly published by the model provider. Search and connected tools can retrieve newer information but do not change the model's built-in cutoff.
- Input price
- $4 / 1MCurrent standard list price per million input tokens unless noted.
- Output price
- $20 / 1MCurrent standard list price per million output tokens unless noted.
- Inputs
- Text, imageMedia types accepted by the listed model endpoint.
- Tools & agents
- Functions, web/file search, code, computer, MCPSelected native tools and agent-building capabilities, not an exhaustive list.
Pricing & comparisons
Estimate your cost
Set your usage. Your estimate updates as you type.
Example: 6,000 input tokens for 10 pages, plus a 500-token summary. Page lengths vary; adjust the numbers below.
One run sends your input to the model once and receives an answer. Tokens are pieces of text: input is what you send, output is the answer you receive.
Advanced options
Daybreak Blue
OpenAI
Estimated total (USD)
For the usage above · USD · API pricing, not a subscription
- Input tokens
- $4.00 per 1M tokens
- Output tokens
- $20.00 per 1M tokens
How this estimate works
Estimates exclude taxes, tools, cache storage/writes, free allowances and custom discounts. Image estimates cover output only, not prompt or reference-image charges. Quality modes differ by model. Unlisted settings are not treated as free.
API and provider access
Where to get Daybreak Blue
OpenAIDaybreak Blue model and API docsResponses API only; needs Daybreak approval and an enabled, non-default API project.OpenAIApply for Daybreak accessIndividuals apply with identity verification; organizations use OpenAI's enterprise request form.OpenAIDaybreak setup in CodexApproved Codex users turn on the Daybreak toggle, or select the alias when signed in with an API key.Amazon BedrockAmazon Bedrock model cardFor eligible customers after Trusted Access enrollment; billed through AWS.
Availability
Regions and access stage
Approval required: available only to individuals and organizations approved for OpenAI's Daybreak program, through the Responses API and Codex.
Access applies only to the approved identity, workspace or API project. Eligible customers can also use it on Amazon Bedrock in US East (Ohio).
Check live availabilityData and training
The route matters.
OpenAI says API inputs and outputs are not used for model training by default. Daybreak approval does not include Zero Data Retention; default abuse-monitoring logs may be kept for up to 30 days unless other controls are approved.
This is a concise reading of the cited provider material, not legal advice. A third-party gateway can have different storage, routing, training, and residency terms from the model maker's direct API.
Read the provider policyFrequently asked questions
Daybreak Blue FAQ
What is Daybreak Blue?
OpenAI's approval-gated access to its flagship models for defensive security work, with fewer cybersecurity refusals. Daybreak Blue is an access tier in OpenAI's Daybreak program, not a separately trained model. The API alias currently points to GPT-5.6 Sol, so it shares that model's 1.05M-token context, 128K-token output limit and February 16, 2026 knowledge cutoff. Approved users can also apply Daybreak Blue to GPT-6 Sol and GPT-6 Luna.
When was Daybreak Blue released?
Daybreak Blue was released on August 7, 2026 according to the cited provider materials.
Where can I access Daybreak Blue?
Approval required: available only to individuals and organizations approved for OpenAI's Daybreak program, through the Responses API and Codex. The access routes listed in this guide are OpenAI and Amazon Bedrock.
How much does Daybreak Blue cost?
$4 input · $20 output / 1M tokens. Billed at the rates of the model behind the alias, currently GPT-5.6 Sol, whose promotional pricing runs at least through November 21, 2026. Cached input costs $0.40 and cache writes $5 per 1M tokens. Above 272K input tokens, the whole request uses $8 input and $30 output rates. OpenAI says prices follow the alias to newer models.
Where is Daybreak Blue available?
Approval required: available only to individuals and organizations approved for OpenAI's Daybreak program, through the Responses API and Codex. Access applies only to the approved identity, workspace or API project. Eligible customers can also use it on Amazon Bedrock in US East (Ohio).
Is my Daybreak Blue API data used for training?
OpenAI says API inputs and outputs are not used for model training by default. Daybreak approval does not include Zero Data Retention; default abuse-monitoring logs may be kept for up to 30 days unless other controls are approved. The policy belongs to the provider route and account terms, so verify it again before production use.
Price comparison
One price per model, using the same input/output mix. We weight each API rate by its share of one million total tokens and use the lowest matching reviewed route. This is a comparison rate, not a cost per task or a quality score. Tokenizers, caching, reasoning and long context can change your actual bill.
75% input + 25% output · no cache discounts
- GPT-6 Luna$0.20
OpenAI
- Mistral Small 4$0.2625
Mistral AI
- GPT-5.6 Luna$0.45
OpenAI
- Gemini 3.8 Flash$1.50
Google
- Claude Haiku 4.5$2.00
Anthropic
- GPT-6 Sol$4.00
OpenAI
- Claude Sonnet 5$4.00
Anthropic
- GPT-5.6 Terra$4.50
OpenAI
- Daybreak Blue$8.00
OpenAI
This model
- Claude Opus 5.5$8.00
Anthropic
- GPT-5.6 Sol$8.00
OpenAI
- Claude Opus 5$10.00
Anthropic
- Claude Fable 5$20.00
Anthropic
- GPT-6 Astra$20.00
OpenAI
- Claude Fable 5.1$20.00
Anthropic
Related comparisons
Head-to-head comparisons
Daybreak Blue vs GPT-6 Sol
An OpenAI model for coding projects and assistants that carry out several steps with tools.
Open full comparisonDaybreak Blue vs GPT-6 Astra
OpenAI's most capable model for difficult end-to-end work, combining frontier reasoning with a million-token context window and a broad set of agent tools.
Open full comparisonDaybreak Blue vs GPT-5.6 Sol
OpenAI's flagship general model for difficult coding, analysis, and professional work, with a very large context window and a broad native tool set.
Open full comparison