OpenLuxVision & multimodal
These OpenLux models accept image input alongside text — screenshots, photos, charts, diagrams and UI captures — and are called with the standard OpenAI image_url content parts. Seven models, three price tiers, one endpoint.
Vision unlocks the workflows that pure text models cannot touch: reading dashboards, describing product photos, transcribing whiteboards, QA-testing interfaces and turning screenshots into bug reports. On OpenLux the image travels inside the same chat request, so a vision feature is usually a few lines added to existing chat code.
The spread between the cheapest and most capable vision model is roughly 40x per token, which is why the tier you choose matters. Flash-tier Gemini handles high-volume reading and extraction; GPT-6 Astra, Claude and Gemini Pro tiers take on chart reasoning, UI understanding and tasks where being wrong is expensive.
The picks
7 models from the live catalogue, with the rate each one is metered at.
Gemini 3 Flash (preview)
Cheapest way to read images at volume — extraction and tagging — $0.0368 per 1M tokens (-92% vs official rates).
Gemini 3.8 Flash
Stable high-throughput vision for production features — $0.0551 per 1M tokens (-92% vs official rates).
Gemini 3.1 Pro (preview)
Deep multimodal thinking for charts, diagrams and layouts — $0.1471 per 1M tokens (-92% vs official rates).
GPT-5.6 Sol
Balanced GPT vision for mixed workloads — $0.1838 per 1M tokens (-96% vs official rates).
GPT-6 Astra
OpenAI's flagship vision reasoning — $0.3677 per 1M tokens (-96% vs official rates).
Claude Sonnet 4.5
Screenshot and UI understanding for coding agents — $0.2647 per 1M tokens (-91% vs official rates).
Claude Fable 5.1
Careful visual review for long-form analysis — $1.765 per 1M tokens (-82% vs official rates).
Vision & multimodal questions
Which models on OpenLux accept image input?
GPT-6 Astra, GPT-5.6 Sol, Gemini 3.8 Flash, Gemini 3 Flash preview, Gemini 3.1 Pro preview, Claude Sonnet 4.5 and Claude Fable 5.1.
How do I send an image in a request?
Use the standard OpenAI content-parts shape: a message content array with an image_url entry, which accepts either a public URL or a base64 data URI. The same payload works on every vision model — only the model name changes.
Which vision model is the cheapest?
Gemini 3 Flash preview at $0.0368 per 1M tokens, about 92% below official list pricing. For reading-heavy pipelines it is usually the right default, with a stronger tier reserved for low-confidence cases.
Do I need separate keys for OpenAI, Anthropic and Google vision models?
No. One OpenLux key covers all of them and every request appears on one invoice, so comparing vision models on your own data costs a model-name change rather than a new provider integration.
Compare the picks head-to-head
GPT-6 Astra vs Gemini 3 Flash (preview)
Rates, strengths and which model to pick.
GPT-5.6 Sol vs Gemini 3 Flash (preview)
Rates, strengths and which model to pick.
Gemini 3.8 Flash vs Gemini 3 Flash (preview)
Rates, strengths and which model to pick.
Gemini 3 Flash (preview) vs Claude Sonnet 4.5
Rates, strengths and which model to pick.
Gemini 3 Flash (preview) vs Claude Fable 5.1
Rates, strengths and which model to pick.
Gemini 3 Flash (preview) vs Gemini 3.1 Pro (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.