OpenLuxCheapest text models
Every text model on OpenLux is metered per 1M tokens at published rates, with discounts already applied. These five carry the lowest rates on the platform, ranked by price, with a note on what each is best at.
Running a product on AI usually starts with a volume problem: classification, drafting, summarisation and chatbot traffic all scale linearly with users. The cheapest way to serve that traffic is a low-cost text tier behind the same API surface as the flagship models, so a single code path can later escalate to a stronger model when a request needs it.
All five models below accept the same OpenAI-compatible request on OpenLux and are billed in USD per token. The discounted rates shown are what gets metered on your invoice — there is no separate list price to negotiate and no minimum commitment.
The picks
5 models from the live catalogue, with the rate each one is metered at.
Gemini 3 Flash (preview)
Lowest text rate on the platform — high-concurrency chat and batch jobs — $0.0368 per 1M tokens (-92% vs official rates).
Gemini 3.8 Flash
Stable long-context generalist, a step up in quality for modular apps — $0.0551 per 1M tokens (-92% vs official rates).
DeepSeek V4
Strong Chinese comprehension and step-by-step reasoning at a budget rate — $0.0613 per 1M tokens (-90% vs official rates).
Qwen3 Max
Dependable Chinese and coding output for bilingual products — $0.0784 per 1M tokens (-88% vs official rates).
Gemini 3.1 Pro (preview)
Deep thinking and multimodal input when a task outgrows the flash tier — $0.1471 per 1M tokens (-92% vs official rates).
Cheapest text models questions
What is the cheapest text model on OpenLux?
Gemini 3 Flash preview at $0.0368 per 1M tokens, roughly 92% below official list pricing. Gemini 3.8 Flash ($0.0551) and DeepSeek V4 ($0.0613) are the next lowest rates on the catalogue.
Do cheap models include the same API features as flagship models?
They are called the same way on the same endpoint: streaming, tools and vision are model-dependent capabilities, not pricing tiers. Budget tiers suit high-volume chat and automation; flagship tiers such as GPT-6 Astra and Claude Opus 5 are better for complex reasoning. Moving between them is a one-field change to the model name.
How is usage metered?
Pay as you go per 1M tokens, billed in USD with the discount already included. The console shows the cost of every request in real time, and there is no subscription requirement to use the API.
Can I mix cheap and flagship models in one application?
Yes. Route the easy majority of requests to a budget tier and escalate hard requests to a flagship model — both sit behind the same key, the same endpoint and the same invoice, so there is no second integration to maintain.
Compare the picks head-to-head
Gemini 3.8 Flash vs Gemini 3 Flash (preview)
Rates, strengths and which model to pick.
Gemini 3 Flash (preview) vs Gemini 3.1 Pro (preview)
Rates, strengths and which model to pick.
Gemini 3 Flash (preview) vs DeepSeek V4
Rates, strengths and which model to pick.
Gemini 3.8 Flash vs Gemini 3.1 Pro (preview)
Rates, strengths and which model to pick.
Gemini 3.8 Flash vs DeepSeek V4
Rates, strengths and which model to pick.
Gemini 3.1 Pro (preview) vs DeepSeek V4
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.