OpenAI
Use Codex freely, starting now.
- 99.5%+ SLA
- 90%-99% cache rate
- Alipay payments
- Only 4.5% of official pricing
Live Pricing
Model pricing, as low as 4.5% of official rates
Anthropic
Claude models from 20% of official pricing
Text models
Unit: USD / 1M tokens| Model | Context | Input - official | Output - official | Cache hit - official | Input - AnixAPI | Output - AnixAPI | Cache hit - AnixAPI |
|---|---|---|---|---|---|---|---|
| gpt-5.6-sol | 1.05M | $5.0000 | $30.0000 | $0.5000 | $0.225095.5% off | $1.3500 | $0.0225 |
| gpt-5.6-terra | 1.05M | $2.0000 | $12.0000 | $0.2000 | $0.090095.5% off | $0.5400 | $0.0090 |
| gpt-5.6-luna | 1.05M | $0.2000 | $1.2000 | $0.0200 | $0.009095.5% off | $0.0540 | $0.0009 |
| gpt-5.5 | 1M | $5.0000 | $30.0000 | $0.5000 | $0.225095.5% off | $1.3500 | $0.0225 |
| gpt-5.4 | 1M | $2.5000 | $15.0000 | $0.2500 | $0.112595.5% off | $0.6750 | $0.0113 |
| gpt-5.4-mini | 400K | $0.7500 | $4.5000 | $0.0750 | $0.033895.5% off | $0.2025 | $0.0034 |
Image generation pricing
Unit: USD / image generation| Model | 1K - official | 2K - official | 4K - official | 1K - AnixAPI | 2K - AnixAPI | 4K - AnixAPI | |
|---|---|---|---|---|---|---|---|
| gpt-image-2 | ≈$0.05 | ≈$0.05 | ≈$0.09 | ≈$0.002395.5% off | ≈$0.0023 | ≈$0.0041 |
Text models
| Model | Context | Input - official | Output - official | Cache hit - official | Input - AnixAPI | Output - AnixAPI | Cache hit - AnixAPI |
|---|---|---|---|---|---|---|---|
| claude-fable-5 | 1M | $10.0000 | $50.0000 | $1.0000 | $2.000080% off | $10.0000 | $0.2000 |
| claude-opus-4-8 | 1M | $5.0000 | $25.0000 | $0.5000 | $1.000080% off | $5.0000 | $0.1000 |
| claude-opus-4-7 | 1M | $5.0000 | $25.0000 | $0.5000 | $1.000080% off | $5.0000 | $0.1000 |
| claude-opus-4-6 | 1M | $5.0000 | $25.0000 | $0.5000 | $1.000080% off | $5.0000 | $0.1000 |
| claude-sonnet-4-6 | 1M | $3.0000 | $15.0000 | $0.3000 | $0.600080% off | $3.0000 | $0.0600 |
| claude-sonnet-5 | 1M | $2.0000 | $10.0000 | $0.2000 | $0.400080% off | $2.0000 | $0.0400 |
How to Start
Quick Start
Quick start
Download and install AnixTools for one-click setup without complex configuration.
Comparison
AnixAPI vs. low-quality relays
| Metric | AnixAPIRecommended | Low-quality relay |
|---|---|---|
| Cache hit rate | High (92%-99%)Data comes from real backend aggregates, closest to actual usage scenarios rather than extreme lab benchmark scores. | Low (mostly <90%, sometimes 0) |
| Stability (SLA) | High (multi-level failover, 99.5%+) | Low (often below 95%) |
| Quality control | Official channels only, with traceable enterprise verification | Hard to verify authenticity |
| Time to first token | Premium multi-region IP routing, milliseconds from direct access | High |
| Actual cost | Low | High (poor cache rate and stability can greatly increase token usage) |
Value
Why choose us?
Stable service and responsive support
A unified API entry is backed by Sub2API and multiple CPA resources, with product surfaces that keep availability visible.
Transparent billing logs
The console organizes balance, API keys, usage trends, and invoice records so every cost has a clear source and teams can review spend with less friction.
Pricing as low as 4.5% of official rates
For high-frequency Codex and agent usage, AnixAPI presents usage cost at 4.