Lumin AI Review: Pricing & Comparison
Launched in 2026; original luminai.cn has moved to luminai.cc, and the main domain now requires a proxy
Last verified: 2026-08-15 · Visit official site →
The current state of a new platform: domain migration and availability swings
Lumin AI (originally luminai.cn) is a relay that launched in 2026. Verified in August 2026: the luminai.cn main domain is no longer directly reachable, and the service has moved to luminai.cc (api.luminai.cc is a New API panel). An official 2026-08-02 notice says the “ai main domain can only be opened via proxy” — i.e. mainland direct connect is no longer available. Third-party monitors (lmspeed) show ~0% recent availability, so reliability is a concern.
If you want to trial it, top up a small amount first and test connectivity and success rates yourself before relying on it.
Models and pricing: ¥1=$1, overseas flagships first
Lumin AI bills per token with a credit rate of ¥1 = $1. Third-party monitors show roughly 97 models, mostly overseas:
- Claude Opus tier: about $2-10 (input) / $10-50 (output) per M tokens depending on channel;
- Gemini 3.5 Flash, grok-imagine (image), and others;
- Codex limited-time channel: about $0.4 / $2.4, though community reports flag account-ban risk.
The earlier “¥5 minimum top-up” and “Kiro endpoint as low as ¥2/10M tokens” could not be re-verified against any live channel — check the console for real prices.
Risk notes
As a new platform, Lumin AI’s core risk is sustained operation and domestic availability: the main domain needs a proxy and recent monitored availability is low, so trial with a small top-up before committing. If you want a stable mainland-direct relay instead, consider YKH.AI or TokenRiver, which have operated longer.
Information verified 2026-08-15. Domain migration, availability and pricing are subject to the live luminai.cc console.
Related Reviews
- NanoBanana: 4K-quality image and video generation, exclusive NanoBanana model series, focused on multimedia generation
- TokenRiver: an enterprise-grade AI model gateway, 650+ models, a favorable ¥1=$1 exchange rate, multi-channel automatic failover, redirect target of the former Shiyun Cloud API
- AIAPIpk: a comparison tool that tracks real-time pricing and availability across relays, a competitive-ranking platform
- UniAPI: the OpenAI/Claude/Gemini three families plus multimodal models like Midjourney/Suno/Kling, mainland direct connect with no proxy needed, billed at per-channel discount rates
Quick facts
| Pricing model | Pay-as-you-go, ¥1 = $1 credit; Claude Opus tier around $2-10/$10-50 per M tokens, Codex limited-time tier $0.4/$2.4; the earlier ¥5 minimum top-up and Kiro ¥2/10M claims could not be re-verified |
|---|---|
| Model coverage | Overseas-first lineup (Claude/Gemini etc.), with roughly 97 models visible to monitors (claude-opus-5, gemini-3.5-flash, grok-imagine, codex series) |
| Latency / SLA | Official notice (2026-08-02) says the main domain now needs a proxy; third-party monitors show ~0% recent availability — reliability is questionable |
| Mainland direct connect | Proxy required |
| Best for | Developers |
| Referral program | No public affiliate program found so far. |
Pros
- Credit rate ¥1 = $1, and overseas flagship models (Claude Opus tier $2-10/$10-50) look cheaper than official pricing
- Has low-cost channels like Codex limited-time for cheaply trying some models
- A new platform that's still iterating, ~97 models covering Claude/Gemini and other mainstream overseas models
Cons
- The main domain has required a proxy since 2026-08-02 — mainland direct connect is no longer available; third-party monitors show ~0% recent availability, so stability risk is high
- The earlier ¥5 minimum top-up and Kiro ¥2/10M token claims could not be re-verified — check the live console
- Short operating history; docs and support are less mature than established platforms
Compare more AI API relays
See the full comparison board — filter by price tier, model coverage, and mainland direct-connect status.
Back to the comparison board →