Model Pricing Reference
Current pricing per million tokens ($/M)
| Model | Input ($/M) | Output ($/M) | Reasoning Level | Notes |
|---|---|---|---|---|
Claude 3 Haiku | $0.25 | $1.25 | 中等偏高 | 最平 Anthropic 模型，200K context |
Claude 3.5 Haiku | $0.80 | $4.00 | 中高 | price/quality sweet spot，200K context |
Claude 3.5 Sonnet | $3.00 | $15.00 | 頂級 | 200K context，最強推理能力 |
Perplexity Sonar Pro | $3.00 | $15.00 | 中高偏上 (search-QA 特強) | 200K context；高級 search 另外按次收費 |
GPT-4.1 | $2.00 | $8.00 | 頂級 | 1M context |
Gemini 1.5 Flash | $0.10 | $0.40 | 中等偏高 | 1M context；極快、極平 |
Gemini 1.5 Pro | $2.20 | $8.80 | 頂級 | 2M context |
Qwen (高配 Max) | $0.20-0.60 | $1.6-2.0 | 中高 | 128K–1M context (按型號) |
GLM-4.5 (Reasoning) | $0.35 | $1.55 | 中高偏上 | ~131K context |
DeepSeek (V3.2 / R1) | $0.03-0.30 | $0.11-2.4 | 次頂級 (推理/coding 特強) | context 一般 32K–160K 級 |
Legend
- Models available in sandbox (DeepSeek, Claude Sonnet, Perplexity)
- • Prices are approximate and may vary by region or volume
- • Context window sizes shown are typical maximums
- • All prices in USD per million tokens