OpenRouter 目前提供 22 个大模型 API。
速度基准测试平均吞吐 85 tok/s。
openrouter.ai| 模型 | 速度 | 延迟 | 测试数 |
|---|---|---|---|
379.88 tok/s | 1.22s | 5 | |
inception/mercury-coder-small-beta | 362.38 tok/s | 1.15s | 10 |
openrouter/sonoma-sky-alpha | 219.22 tok/s | 2.18s | 5 |
140.85 tok/s | 2.05s | 10 | |
openrouter/sherlock-dash-alpha | 103.39 tok/s | 1.21s | 10 |
100.35 tok/s | 6.82s | 40 | |
83.51 tok/s | 4.78s | 15 | |
83.19 tok/s | 22.71s | 5 | |
65.69 tok/s | 4.69s | 5 | |
58.08 tok/s | 14.05s | 15 |
| 时间 | 模型 | 速度 | 延迟 |
|---|---|---|---|
| 04/13/2026, 03:04 | z-ai/glm-5.1 | 63.01 tok/s | 12.13s |
| 04/12/2026, 11:56 | openai/gpt-oss-120b:free | 34.72 tok/s | 1.75s |
| 04/11/2026, 04:25 | arcee-ai/trinity-large-preview:free | 5.85 tok/s | 1.51s |
| 04/11/2026, 04:15 | arcee-ai/trinity-large-preview:free | 5.83 tok/s | 1.18s |
| 04/08/2026, 14:33 | z-ai/glm-5.1 | 43.91 tok/s | 14.92s |
A third-party OpenRouter-compatible API relay providing access to multiple AI models.
12 个共享模型Seamee API provides an AI model relay for accessing multiple LLMs through OpenAI-compatible endpoints.
12 个共享模型Provides AI-generated content APIs for various applications, including text and image generation.
11 个共享模型Synapse is an OpenAI-compatible API relay service providing access to multiple AI models with unified endpoints.
10 个共享模型A non-profit AI infrastructure offering free access to integrated large language models with privacy-focused, no-logging policies.
10 个共享模型人人 API provides a unified LLM API gateway with competitive pricing, no subscription required, and access to multiple models.
9 个共享模型