OpenLuxBiggest discounts
OpenLux aggregates upstream capacity and passes the saving through to published rates. These models carry the biggest discounts against official list pricing — the percentage shown is already inside the rate you are billed.
Discounts are not a coupon step or a negotiated enterprise tier: every rate on this page is the metered, pay-as-you-go price on the platform. The discount against official pricing is shown so you can judge what the same workload would cost on direct provider accounts, where a single key covers one provider instead of the full catalogue.
The list is ranked by discount depth. Tooling, context and capability differ between the models, so the right pick still depends on your workload — the discount determines the budget, not the fit.
The picks
6 models from the live catalogue, with the rate each one is metered at.
GPT-6 Astra
The catalogue's flagship reasoner with the deepest discount — $0.3677 per 1M tokens (-96% vs official rates).
GPT-5.6 Sol
Value-tier GPT-5.6 for production volume at the same discount depth — $0.1838 per 1M tokens (-96% vs official rates).
GPT Image 2 C
Image generation and editing with the largest saving in the image category — $0.00882 per request (-93% vs official rates).
Gemini 3.8 Flash
High-throughput multimodal workhorse from the Gemini line — $0.0551 per 1M tokens (-92% vs official rates).
Gemini 3 Flash (preview)
The cheapest entry point on the platform, also heavily discounted — $0.0368 per 1M tokens (-92% vs official rates).
Gemini 3.1 Pro (preview)
Multimodal thinking tier for hard requests that still need a low rate — $0.1471 per 1M tokens (-92% vs official rates).
Biggest discounts questions
How large are the biggest discounts on OpenLux?
Up to 96% against official list pricing on GPT-6 Astra and GPT-5.6 Sol, with most flagship models sitting between 82% and 96%. The discounted rate is what the platform meters per request.
Do discounts change how requests are routed or answered?
No. The model name you pass is the model that answers — pricing reflects aggregated upstream capacity, not model substitution or a downgraded tier. Smart routing only retries across regions when a node fails.
Do the discounts apply to pay-as-you-go billing?
Yes. The rates listed here and on the model pages are already discounted, whether you use pay-as-you-go top-ups or a monthly or yearly plan with credit included.
Which discounted models give the best capability per dollar?
GPT-5.6 Sol and Gemini 3.8 Flash for production chat at scale, GPT-6 Astra when reasoning quality matters, and gpt-image-2-c for image pipelines — all at discount depths above 90%.
Compare the picks head-to-head
GPT-6 Astra vs GPT-5.6 Sol
Rates, strengths and which model to pick.
GPT-6 Astra vs Gemini 3.8 Flash
Rates, strengths and which model to pick.
GPT-6 Astra vs Gemini 3 Flash (preview)
Rates, strengths and which model to pick.
GPT-6 Astra vs Gemini 3.1 Pro (preview)
Rates, strengths and which model to pick.
GPT-5.6 Sol vs Gemini 3.8 Flash
Rates, strengths and which model to pick.
GPT-5.6 Sol vs Gemini 3 Flash (preview)
Rates, strengths and which model to pick.
Put these picks behind one key
OpenLux meters every model in this collection on the same OpenAI-compatible endpoint, with pay-as-you-go billing and free trial credit to start.