Twinkle Model
PaidRoute compatible AI API clients through a third-party gateway with usage controls
About
Twinkle Model is an independent, third-party AI API gateway at big-model.smart-agi.com. Its public interface lists multiple model routes and usage-based prices. It is not an official OpenAI, Anthropic or Google service; upstream authorization, model authenticity and performance have not been independently verified.
✨ Key Features
- Connect compatible clients by changing the base URL and API key.
- Use advertised Anthropic Messages, OpenAI Chat/Responses or Gemini-format endpoints, depending on configured routes.
- Request images or asynchronous video jobs where the selected group supports them.
- Create separate keys for projects or members with spending caps and expiry dates.
- Apply IP allow/deny rules and per-key concurrency controls.
- Review request-level model, token and charge logs and export usage.
- Filter the public model plaza by platform, group or model name.
- Compare group-specific input, output and cached-token pricing where published.
💳 Pricing
- Pay-as-you-go — Prepaid balance; rates vary by model and routing group. The public plaza covers token-priced text routes, per-request items and per-second video jobs, rather than a single subscription tariff.
- A published example route labeled
claude-haiku-4-5shows CNY0.15 input and CNY0.75 output per million tokens in the cheapest listed group. This is the gateway's quote, not a claim about official provider pricing or model authenticity. - Listed video routes span CNY0.30–0.45/second; a low-cost text route is listed at CNY0.007/request. Cached-token charges and other routes differ.
- Displayed CNY costs incorporate recharge and group multipliers. Missing prices appear as a dash, not free usage. Check the complete live model plaza before funding an account; no permanent free tier was verified.
⚖️ Alternatives Compared
| Product | Best for | Public pricing | Key difference |
|---|---|---|---|
| Twinkle Model | Compatible-client routing | Group-dependent usage rates | Independent gateway and CNY price display |
| OpenRouter | Multi-provider model APIs | Free and usage-based paid plans | Provider selection and routing platform |
| Replicate | Running hosted models | Usage-based | Model deployment and hardware options |
🎯 Use Cases
- Developers: evaluate an alternate endpoint in a compatible client.
- Project maintainers: separate API keys and spending limits.
- Administrators: restrict key use to selected IP addresses.
- Finance operators: export request charges for reconciliation.
- Multimedia developers: assess supported asynchronous image/video routes.
Twinkle Model Website Overview
Twinkle Model has an Ahrefs Domain Rating of 2.1 and falls within Cloudflare Radar's global top 200,000 popularity bucket. This bucket is unordered, and neither signal represents actual visits.
- Domain rating
- 2.1
- Popularity position
- Top 200,000
- Primary category
- Unavailable
- Ranked locations
- 0
Ahrefs
Cloudflare Radar
Cloudflare Radar
Cloudflare Radar
Website categories
No category data is available yet.
Top locations by rank
No location ranking is available yet.
Data coverage
Free popularity and domain-authority signals only; no estimated visits or engagement metrics are shown.
Sources: Domain Rating by Ahrefs · Cloudflare Radar · Nov 21, 2026. Cloudflare Radar data is licensed under CC BY-NC 4.0.
Similar tools
BetterToken
PaidA prepaid multi-model API gateway for compatible coding tools and applications
Hugging Face Spaces
FreemiumGit-based hosting for interactive ML demos and model-powered apps on the Hugging Face Hub
you.bot
PaidUnified AI model API gateway for image, video, audio and text with usage-based credits
Google AI Studio
FreemiumGoogle’s browser workspace for prototyping with Gemini and building multimodal AI applications
