Grok Imagine Image
Fast creative image generation, prompt variations and reference-image workflows.
Open hubEnglish home / Model hubs
English model directory
Model hubs built around real API jobs
Each hub gives you the model ID, use cases, billing boundaries, a minimal request and an acceptance checklist. Use the dashboard for live availability and final billing.
Image APIs
Gemini 3.1 Flash Image
Multimodal image generation and editing for product, social and batch creative work.
Open hubVideo APIs
Chinese model APIs
DeepSeek V4 Pro
Reasoning, coding and long-context tasks through an OpenAI-compatible request shape.
Open hubQwen 3.7 Max
Multilingual reasoning, structured outputs and tool-oriented workloads.
Open hubGLM-5.3
Chinese-language workflows, coding and structured task automation.
Open hubKimi K3
Long-context reading, research notes and Chinese-English research workflows.
Open hubChoose a starting point
Use the hub that matches the output you need to ship, then run the same acceptance test on two candidates. The model name is only a starting hypothesis; quality, latency, failure behavior and ledger cost decide whether it belongs in production.
| Need to deliver | Start with | First comparison |
|---|---|---|
| Product or social images | Grok Imagine Image or Gemini 3.1 Flash Image | Prompt adherence, usable-asset rate and cost per accepted image |
| Short video clips | Seedance 2.5 or Grok Imagine Video | Terminal success, duration, download reliability and final task charge |
| Chinese reasoning or code | DeepSeek V4 Pro | Fixed test set, engineering acceptance and input/output usage |
| Structured multilingual output | Qwen 3.7 Max or GLM-5.3 | Schema validity, retry rate and language consistency |
| Document-heavy research | Kimi K3 | Constraint coverage, source fidelity and context cost |
What every hub contains
Each page is written as a deployment decision record: exact model ID, supported workflow assumptions, billing unit to verify, a minimal request, failure boundaries, an acceptance checklist and questions that commonly cause duplicate requests or misleading cost comparisons.
Availability and final billing are always controlled by the live dashboard. Public pages explain how to test a route; they do not replace provider documentation or promise a fixed quality, rate or latency.