Haiflux is a high-performance, privacy-first API gateway providing global access to top-tier Eastern LLMs, completely bypassing regional KYC restrictions.
| Model Name | Input Price (1M) | Output Price (1M) | Best Use Case & Benchmarks |
|---|---|---|---|
| deepseek-r1 | $2.00 | $6.50 | Delivers OpenAI o1-level advanced reasoning. Utilizes chain-of-thought for complex mathematics and logic. |
| deepseek-v4-pro | $3.00 | $6.50 | Direct alternative to GPT-4 Turbo. State-of-the-art general intelligence for enterprise tasks. |
| deepseek-v4-flash | $0.30 | $0.80 | Outperforms GPT-4o-mini. Lightning-fast inference built for high-volume, real-time processing. |
| qwen3.7-max | $4.50 | $11.00 | Matches GPT-4o and Claude 3.5 Sonnet. The ultimate powerhouse for complex agentic workflows. |
| qwen3-coder-plus | $1.50 | $5.50 | A coding powerhouse benchmarking against Claude 3.5 Sonnet. Engineered for seamless IDE integration. |
| qwen-long-latest | $0.30 | $0.80 | Rivals Gemini 1.5 Pro's massive context window. Flawlessly ingests millions of tokens for document parsing. |
| qwen3-vl-plus | $0.70 | $4.00 | Matches GPT-4o Vision capabilities. Accurately understands complex UI layouts and charts. |
| qwen3.6-flash | $0.60 | $2.60 | The ultimate Llama-3-70B killer. Optimized for extreme speed and token efficiency. |
| glm-5.1 | $2.00 | $7.00 | Competes directly with Gemini 1.5 Pro. Exceptional RAG capabilities and precise tool execution. |
Primary Base URL: https://api-haiflux.com/v1
Alternative/Redundant Base URL: https://api-tide.com/v1
Get started anonymously at: https://haiflux.com