模型市场
按价格、缓存命中、可用性与首字延迟对比 GPT、Claude、Gemini、DeepSeek、GLM 等线路。
| 模型 | 供应商 | 上下文 | 输入 | 输出 | 缓存 | 可用性 | 首字 | 调用量 |
|---|---|---|---|---|---|---|---|---|
| gemini-3-flash Gemini 3 Flash | viapi | 1000K | $0.1200 | $0.4500 | $0.0200 | 98.4% | 610ms | 22.2K |
| deepseek-v3.2 DeepSeek V3.2 | 森林AI | 128K | $0.1400 | $0.2800 | $0.0200 | 99.1% | 890ms | 88.4K |
| gpt-4o-mini GPT-4o mini | TokenX | 128K | $0.1500 | $0.6000 | $0.0200 | 99.2% | 430ms | 54.0K |
| llama-4-maverick Llama 4 Maverick | Kriora | 256K | $0.1800 | $0.5500 | $0.0200 | 95.1% | 720ms | 8.0K |
| qwen3.5-plus Qwen 3.5 Plus | Kriora | 128K | $0.2000 | $0.8000 | $0.0300 | 96.8% | 640ms | 13.0K |
| glm-4.6 GLM-4.6 | viapi | 200K | $0.2200 | $0.8800 | $0.0300 | 97.6% | 980ms | 15.4K |
| claude-sonnet-4-6 Claude Sonnet 4.6 | TokenX | 200K | $0.2760 | $1.38 | $0.0276 | 98.8% | 2100ms | 329 |
| gpt-5.6-terra GPT-5.6 Terra | viapi | 272K | $0.4600 | $2.75 | $0.0500 | 96.2% | 4200ms | 1.4K |
| gpt-5.5 GPT-5.5 | 森林AI | 272K | $0.8000 | $4.80 | $0.0800 | 94.0% | 11127ms | 339.2K |
| claude-opus-4-8 Claude Opus 4.8 | 森林AI | 1000K | $0.9800 | $4.90 | $0.1000 | 99.5% | 3491ms | 6.1K |
| gemini-2.5-pro Gemini 2.5 Pro Official | 官方 Gemini | 1000K | $1.25 | $10.00 | $0.3100 | 99.4% | 690ms | 7.0K |
| gpt-4o GPT-4o Official | 官方 OpenAI | 128K | $2.50 | $10.00 | $1.25 | 99.8% | 520ms | 12.0K |
| claude-sonnet-4 Claude Sonnet 4 Official | 官方 Anthropic | 200K | $3.00 | $15.00 | $0.3000 | 99.7% | 780ms | 9.9K |