Model profile
GPT-5.3 Codex Spark
MaintenanceLightweight fast variant of GPT-5.3 Codex for rapid coding tasks.
Everything you need to evaluate this model before wiring it into production.
This model remains in the public catalog, but Token101 has paused live routing while upstream availability is being restored.
gpt-5.3-codex-sparkProvider
OpenAI
Launch status
Maintenance
Context window
1.0M
Prompt caching
Supported
API routing is pausedView Pricing
Capabilities
Check context size, supported modalities, and output limits before you integrate.
Context window
1.0M
Max output
16.4K
Input modalities
Text, Image
Output modalities
Text
Prompt caching
Supported