Not every task needs your most expensive model.
Simplex Brain is the router at the center of the GoSimplex platform — reading each task and deciding which model and which path actually fits it, instead of defaulting every call to the priciest option out of habit. Start on pay as you go in minutes: sign in, choose Brain and create a key. For larger spend, tell us about your workload and we will set up share of savings with you.
Right-sized routing, not one model for everything.
Most AI spend goes to calls that never needed the biggest model in the first place. Simplex Brain reads what a task actually requires and routes it accordingly — a lightweight model for a simple lookup, a heavier one only when the task calls for it.
Illustrative of results seen in certain use cases — not an audited benchmark, and not a guarantee for every workload. Actual savings depend on the mix of tasks being routed.
You pay the model providers at cost, plus a fee that fits the size of your spend.
Small teams pay a simple percentage on top of the provider bill. Larger teams pay a share of what Brain saves them. On every request Brain returns what the model it chose cost and what the full request would have cost on a strong default model, so the saving is visible call by call.
Pay as you go
- •Billed on what you use, no commitment, no baseline to agree
- •A good start for small or variable use
Share of savings
- •Model providers are billed at what they charged
- •Plus 25% of the verified savings against a strong-model baseline
- •If Brain saves nothing, you pay only the monthly minimum
Enterprise
- •Your own baseline model
- •Service agreement and volume terms
- •Scoped together with your team
What a month looks like for you
Share of savings. Without Brain, your AI calls would cost ₹1,00,000 a month. With Brain, the model providers bill you ₹41,000. Brain's fee is 25% of the ₹59,000 saved, which is ₹14,750. You pay ₹55,750 in all and keep ₹44,250 every month.
Pay as you go. If the model providers bill you ₹10,000 in a month, Brain adds 8%, which is ₹800. You pay ₹10,800 in all.
Illustrative, based on a 59% saving seen on a test set of 245 tasks. Actual savings depend on your mix of tasks.
Terms of the price
- Which plan applies
- Pay as you go has no monthly limit other than your wallet balance: you top up, calls draw from it, and we email you when it runs low. Share of savings suits larger or steadier use. You can start pay as you go yourself. For share of savings we agree the terms with you.
- How savings are measured
- Per request, Brain compares the cost of the model it chose with the cost of the full request on a strong default model, net of Brain's own overhead.
- What you call
- OpenAI-compatible chat completions, a model list and embeddings.
- Default limits
- Up to 8,192 output tokens per call, requests up to 10 MB, and 64 inputs per embeddings call.
- Access
- Pay as you go: sign in to the GoSimplex console, choose Brain and create your key. Share of savings: set up by us. Talk to us and we will agree the baseline.
- Billing
- Monthly. Prices are in Indian rupees and exclude 18% GST.
The router at the center of the platform.
Simplex Brain decides which model and which path handles each task. Migrus, Nyaya, and Conexon are what make that decision safe to trust. Vak is the part your business actually meets.