DeepSeek V3.2
DeepSeek
GPT-4-class general model. The most cost-effective model in the world for code & chat.
- 上下文
- 64K
- 类型
- 对话
- 输入
- $0.14/M
- 输出
- $0.28/M
模型
同样的 OpenAI SDK,同样的 `model` 字段。智能路由自动挑选最便宜的健康通道。下方列出网络中的全部模型,含实际价格、上下文窗口、能力支持。点击任意卡片复制 model id。
DeepSeek
GPT-4-class general model. The most cost-effective model in the world for code & chat.
DeepSeek
Open reasoning model rivalling o1 on math & logic, at 1/30 the price.
Moonshot
Long-context tool-use champion. Strong agentic behaviour, free 128K context.
Zhipu
Coder-tuned. Strong on Chinese function calling and tool routing.
Alibaba
256K context. Strong on multilingual and bilingual reasoning.
Alibaba
Coder-specialised Qwen variant. SOTA on HumanEval-CN.
ByteDance
Strong on image understanding & Chinese OCR. Mature production model.
MiniMax
Fastest Chinese chat model — 130 tokens/sec on cold start.
Anthropic
Frontier reasoning. Pass-through pricing — exactly the official rate.
Anthropic
Workhorse Anthropic model. 5× cheaper than Opus, ~90% the quality.
OpenAI
Multimodal frontier. Pass-through pricing.
OpenAI
Latest GPT generation. Pass-through pricing.
OpenAI
OpenAI flagship reasoning. Pass-through pricing.
2M token context. Pass-through pricing.
Kling
Best Chinese text-to-video. 5s/10s clips, 720p/1080p output.
国产模型通常在上游发布 7 天内上架。在 GitHub 提一个 issue,我们优先排期。