The Complete Guide to GPT-5.6 Mini / Nano
The smallest, cheapest tiers beyond Sol/Terra/Luna
Beyond the officially-promoted Sol, Terra, and Luna tiers, the GPT-5.6 family also includes smaller, cheaper Mini and Nano tiers; QCode sees 5 and 2 real calls respectively in 30 days — thin, but genuinely on sale.
Four key facts
Mini's 30-day usage
gpt-5.6-mini's real call count on QCode in the last 30 days — thin, but genuinely positive, proving it's routable.
Nano's 30-day usage
gpt-5.6-nano's 30-day call count — among the thinnest of any sellable model on the platform.
Positioning
Mini and Nano are smaller, cheaper tiers within the GPT-5.6 family beyond Sol/Terra/Luna; our earlier GPT-5.6 page only covered those three flagship tiers.
Use case
Suited to extremely cost-sensitive scenarios where the task itself is simple — like very high-frequency classification or basic formatting.
What Mini and Nano are
OpenAI publicly promotes the GPT-5.6 family's Sol, Terra, and Luna tiers (already covered in depth on this site), but the family also includes smaller, cheaper Mini and Nano tiers — following the GPT series' longstanding 'mini/nano' tradition of a smaller parameter footprint in exchange for lower unit cost, suited for extremely price-sensitive, simple tasks. QCode's pricing table lists both gpt-5.6-mini and gpt-5.6-nano, and both have real call records — meaning they're genuinely routable, not just listed on a price sheet.
Why these two tiers never got their own page before
Our earlier GPT-5.6 coverage focused on the officially-promoted Sol, Terra, and Luna tiers (previewed 2026-06-26, generally available 07-09); Mini and Nano have gotten noticeably less public coverage and discussion — an 'on sale, but nobody's written it up' tier. This page re-verifies against real 30-day usage data: both are genuinely in QCode's sellable lineup, just with much thinner call volume than the flagship tiers.
Timeline
The GPT-5.6 family previews, followed by general availability on July 9; official messaging focuses on Sol, Terra, and Luna.
Mini and Nano launch alongside as smaller, cheaper tiers, without the same level of official promotion.
QCode's 30-day CRS usage data confirms both gpt-5.6-mini (5 calls) and gpt-5.6-nano (2 calls) have real, routable usage.
Confirmed vs. needs your own verification
Confirmed
Both gpt-5.6-mini and gpt-5.6-nano are in QCode's pricing table, and 30-day CRS usage pulled 2026-08-27 shows 5 and 2 calls respectively — proving they're genuinely routable.
Needs your own verification
Given how thin the usage is, public benchmark data on their exact capability boundaries (like complex reasoning performance) is limited — test at small scale for your specific use case rather than assuming they behave like Sol/Terra/Luna.
Good fit for Mini/Nano vs. needs a higher tier
A good fit for Mini/Nano
Extremely simple tasks, extreme cost sensitivity, and an acceptable lower capability ceiling — like very-high-volume simple classification or formatting.
Needs Sol/Terra/Luna or higher
Complex reasoning, long-context comprehension, or high-quality generation — Mini/Nano's ceiling isn't built for this; use a flagship tier instead.
How to use them
Call gpt-5.6-mini or gpt-5.6-nano directly by model ID on QCode, with the same interface as other GPT models. Given how thin the usage data is, run a small-scale test to verify output quality before committing to large production batches.
What to do on QCode
gpt-5.6-mini and gpt-5.6-nano share one key with the whole GPT lineup on QCode, billed at official price × service rate. For extremely cost-sensitive lightweight tasks, test with these tiers first before deciding whether to step up to Sol/Terra/Luna.
FAQ
Can Mini and Nano really be called right now?
Yes. QCode's 30-day CRS usage data shows real calls to both (5 for Mini, 2 for Nano), confirming they're routable, not just listed on a price sheet.
Why is usage so thin?
Partly because these tiers target extremely lightweight use cases with inherently lower demand, and partly because they haven't gotten the same official promotion as Sol/Terra/Luna, so awareness is lower.
How do Mini/Nano relate to Sol/Terra/Luna?
All are part of the GPT-5.6 family; Sol/Terra/Luna are the officially-promoted flagship-to-mid tiers, while Mini/Nano are smaller, cheaper supplementary tiers.
What are Mini/Nano good for?
Extremely cost-sensitive scenarios where the task itself is simple — high-frequency simple classification or formatting; not recommended for complex reasoning tasks.
How do I decide between Mini and Nano?
They're positioned similarly, with Nano usually smaller and cheaper than Mini; test both at small scale on your specific task before deciding.
How are these two tiers billed on QCode?
At official price × service rate, matching the billing logic for other GPT-series models, with no additional markup.
Sources
The GPT-5.6 family's preview and general-availability dates (2026-06-26, 07-09) come from an already-verified GPT-5.6 page on this site; gpt-5.6-mini and gpt-5.6-nano's sale status and 30-day usage figures come from QCode platform usage verification, pulled 2026-08-27.
Extremely lightweight tasks? Try Mini/Nano first
One QCode key calls the whole GPT-5.6 lineup, including Mini and Nano, billed at official price × service rate.
Related reading
The complete GPT-5.6 guide (Sol/Terra/Luna)
Detailed benchmarks and pricing for the GPT-5.6 family's flagship tiers.
The complete QCode.cc pricing guide
The 7 plan tiers and how billing works.
Subscription vs. API: will a 24/7 agent eat through your plan?
Cost-optimization thinking for extremely lightweight tasks.
This page's model positioning is based on public information and QCode's platform usage verification, and is not an official OpenAI statement; actual call volume and sale status are governed by live QCode platform data.