OpenAI Flex processing is winding down, with no GPT-6.2 support
The half-price Flex tier stays on older models for now. On GPT-6.2, Batch and background windows are the discounted options.
OpenAI is winding down Flex processing, the beta service tier that bills Responses and Chat Completions requests at Batch API rates in exchange for slower responses. OpenAI's Flex processing guide, checked on November 21, 2026, says Flex is unavailable on GPT-6.2 Astra and GPT-6.2 Sol, that it won't be added to future models, and that a retirement date will follow once confirmed. The November 17 entry in OpenAI's API changelog now carries the same note.
What changed
- Requests with
service_tier: "flex"keep working on models that already support Flex while the beta winds down. - Flex is not available on
gpt-6.2-astraorgpt-6.2-sol, and OpenAI says it won't add Flex to new models. - No retirement date yet. OpenAI says it will publish the date and migration guidance in the Flex processing guide.
- Flex tokens cost the same as Batch API tokens, half of Standard on the GPT-6 models.
gpt-6-luna, for example, is $0.05 input and $0.25 output per million tokens on Flex, against $0.10 and $0.50 on Standard. - For work that can wait, OpenAI points to background mode in the Responses API: a
2hcompletion window costs 30% below Standard and an8hwindow 40% below, ongpt-6.2-astra,gpt-6.2-sol,gpt-6.1-terraandgpt-6-luna(our November 17 coverage).
Flex prices today
Short-context prices per million tokens from OpenAI's pricing page on November 21, 2026. OpenAI lists Flex on these models only while the beta winds down.
| Model | Standard input | Standard output | Flex input | Flex output |
|---|---|---|---|---|
gpt-6-astra | $10.00 | $50.00 | $5.00 | $25.00 |
gpt-6.1-sol | $2.00 | $10.00 | $1.00 | $5.00 |
gpt-6.1-terra | $1.00 | $5.00 | $0.50 | $2.50 |
gpt-6-sol | $2.00 | $10.00 | $1.00 | $5.00 |
gpt-6-luna | $0.10 | $0.50 | $0.05 | $0.25 |
gpt-5.6-sol | $4.00 | $20.00 | $2.00 | $10.00 |
gpt-5.6-terra | $2.00 | $12.00 | $1.00 | $6.00 |
gpt-5.6-luna | $0.20 | $1.20 | $0.10 | $0.60 |
What it means on GPT-6.2
A team that runs bulk jobs on Flex to halve its bill loses that option when it moves to GPT-6.2. The discounted paths left on GPT-6.2 are the Batch API and the background windows. For gpt-6.2-sol, per million tokens, from OpenAI's pricing page (background rates computed from OpenAI's stated 30% and 40% discounts):
| Option | Input | Output |
|---|---|---|
| Standard | $2.00 | $10.00 |
Background 2h window | $1.40 | $7.00 |
Background 8h window | $1.20 | $6.00 |
| Batch API | $1.00 | $5.00 |
| Flex | Not available | Not available |
Batch is still the cheapest, but it uses a separate job flow. The background windows keep the normal Responses API request flow, need service_tier set to default and don't support streaming.
What to do
Find requests that set service_tier: "flex" before you switch them to a GPT-6.2 model. For work that can wait a few hours, move them to a background window:
{
"model": "gpt-6.2-sol",
"background": true,
"service_tier": "default",
"background_options": { "completion_window": "8h" },
"input": "..."
}On RouterPlus
Flex, Batch and background windows are OpenAI options on requests sent to OpenAI's API. The RouterPlus catalog lists standard per-token rates. Per million tokens, from the catalog on November 21, 2026:
| Model | Provider | Input | Output |
|---|---|---|---|
| GPT-6.2 Astra | OpenAI (direct API) | $8.00 | $40.00 |
| GPT-6.2 Sol | OpenAI (direct API) | $2.00 | $10.00 |
| GPT-6 Astra | OpenAI (direct API) | $10.00 | $50.00 |
| GPT-6 Astra | Inference.net | $10.00 | $50.00 |
| GPT-6.1 Sol | OpenAI (direct API) | $2.00 | $10.00 |
| GPT-6 Luna | OpenAI (direct API) | $0.10 | $0.50 |
| GPT-6 Luna | Inference.net | $0.10 | $0.50 |
See all OpenAI models on RouterPlus.
