500,000+ checks completed with zero privacy leaks
For account security, we recommend using a test-only KEY. We take account security as seriously as you do
| Site | Channel Name | Price | Trend | Token usage | Uptime | Fake % | Latency | Status |
|---|---|---|---|---|---|---|---|---|
数智AIBiz | gpt plus | 4 | 0% | Higher | 100% | Barely | 6.9s | |
MuskAIBiz | Codex-企业级 | 2 | +14% | — | 100% | Barely | 5.9s | |
SudoCodeBiz | default | 1 | 0% | Much lower | 100% | Barely | 7s | |
ModConBiz | codex-pro-官池 | 1.5 | +67% | Higher | 98% | Barely | 7.5s | |
LLMAPIBiz | default | 5 | 0% | — | 98% | Barely | 3.1s | |
YKH.AIBiz | 大客户专用 | 2.5 | 0% | Higher | 100% | Barely | 7.3s |
SiteChannel Name | PriceTrend | UptimeFake % | LatencyToken usage | Status |
|---|---|---|---|---|
数智AIBiz gpt plus | 40% | 100%Barely | 6.9sHigher | |
MuskAIBiz Codex-企业级 | 2+14% | 100%Barely | 5.9s— | |
SudoCodeBiz default | 10% | 100%Barely | 7sMuch lower | |
ModConBiz codex-pro-官池 | 1.5+67% | 98%Barely | 7.5sHigher | |
LLMAPIBiz default | 50% | 98%Barely | 3.1s— | |
YKH.AIBiz 大客户专用 | 2.50% | 100%Barely | 7.3sHigher |
| Site | Channel Name | Price | Trend | Token usage | Uptime | Fake % | Latency | Status |
|---|---|---|---|---|---|---|---|---|
Green Mountain青山Biz | CC官转max自营 可外接 | 6 | 0% | Average | 100% | Barely | 3.4s | |
DasuApiID | claude max | 3.8 | -43% | Average | 100% | Barely | 4.1s | |
我的刀盾Biz | cc-max | 4 | -33% | Average | 100% | Barely | 3.5s | |
词易Biz | cc-max-满血 | 8 | 0% | Higher | 100% | Barely | 3.4s | |
RunAPIBiz | coding | 10.2 | -3.1% | Average | 100% | Low | 3.2s | |
PackyCodeBiz | cc | 10 | 0% | Average | 100% | Low | 3s |
SiteChannel Name | PriceTrend | UptimeFake % | LatencyToken usage | Status |
|---|---|---|---|---|
Green Mountain青山Biz CC官转max自营 可外接 | 60% | 100%Barely | 3.4sAverage | |
DasuApiID claude max | 3.8-43% | 100%Barely | 4.1sAverage | |
我的刀盾Biz cc-max | 4-33% | 100%Barely | 3.5sAverage | |
词易Biz cc-max-满血 | 80% | 100%Barely | 3.4sHigher | |
RunAPIBiz coding | 10.2-3.1% | 100%Low | 3.2sAverage | |
PackyCodeBiz cc | 100% | 100%Low | 3sAverage |
| Site | Channel Name | Price | Trend | Token usage | Uptime | Fake % | Latency | Status |
|---|---|---|---|---|---|---|---|---|
RunAPIBiz | coding | 10.2 | -3.1% | Average | 100% | Barely | 3.6s | |
DasuApiID | claude max | 3.8 | -43% | Average | 100% | Barely | 4.2s | |
PackyCodeBiz | cc | 10 | 0% | Average | 100% | Barely | 3.3s | |
Green Mountain青山Biz | CC官转max自营 可外接 | 6 | 0% | Average | 98% | Barely | 4.4s | |
UU APIBiz | CC满血MAX -限cc内部 | 9 | -10% | Average | 100% | Barely | 4.1s | |
我的刀盾Biz | cc-max | 4 | -33% | Average | 98% | Barely | 3.7s |
SiteChannel Name | PriceTrend | UptimeFake % | LatencyToken usage | Status |
|---|---|---|---|---|
RunAPIBiz coding | 10.2-3.1% | 100%Barely | 3.6sAverage | |
DasuApiID claude max | 3.8-43% | 100%Barely | 4.2sAverage | |
PackyCodeBiz cc | 100% | 100%Barely | 3.3sAverage | |
Green Mountain青山Biz CC官转max自营 可外接 | 60% | 98%Barely | 4.4sAverage | |
UU APIBiz CC满血MAX -限cc内部 | 9-10% | 100%Barely | 4.1sAverage | |
我的刀盾Biz cc-max | 4-33% | 98%Barely | 3.7sAverage |
| Site | Channel Name | Price | Trend | Token usage | Uptime | Fake % | Latency | Status |
|---|---|---|---|---|---|---|---|---|
RunAPIBiz | claude_normal | 10.2 | -3.1% | Higher | 100% | Barely | 3.2s | |
PackyCodeBiz | cc | 10 | 0% | Average | 100% | Barely | 3.3s | |
Green Mountain青山Biz | CC官转max自营 可外接 | 6 | 0% | Average | 100% | Barely | 3.4s | |
9527codeBiz | claude | 7.5 | 0% | Higher | 100% | Barely | 4.6s | |
ModConBiz | cc-max-满血官池 | 8.5 | 0% | Higher | 100% | Barely | 4.5s | |
CodeRelayBiz | ClaudeCode-Max | 7.5 | 0% | Average | 100% | Barely | 3.1s |
SiteChannel Name | PriceTrend | UptimeFake % | LatencyToken usage | Status |
|---|---|---|---|---|
RunAPIBiz claude_normal | 10.2-3.1% | 100%Barely | 3.2sHigher | |
PackyCodeBiz cc | 100% | 100%Barely | 3.3sAverage | |
Green Mountain青山Biz CC官转max自营 可外接 | 60% | 100%Barely | 3.4sAverage | |
9527codeBiz claude | 7.50% | 100%Barely | 4.6sHigher | |
ModConBiz cc-max-满血官池 | 8.50% | 100%Barely | 4.5sHigher | |
CodeRelayBiz ClaudeCode-Max | 7.50% | 100%Barely | 3.1sAverage |
| Site | Channel Name | Price | Trend | Token usage | Uptime | Fake % | Latency | Status |
|---|---|---|---|---|---|---|---|---|
PackyCodeBiz | cc | 6 | 0% | Average | 100% | Barely | 2.8s | |
RunAPIBiz | coding | 6.1 | -3.1% | Average | 100% | Barely | 3s | |
SudoCodeBiz | default | 6 | 0% | Average | 100% | Barely | 3.1s | |
9527codeBiz | claude | 4.5 | 0% | Higher | 100% | Barely | 4.8s | |
灵算Biz | cc-max-满血 | 4.8 | +6.7% | Much higher | 100% | Barely | 3.5s | |
UU APIBiz | CC满血MAX -限cc内部 | 5.4 | -10% | Average | 100% | Barely | 3.3s |
SiteChannel Name | PriceTrend | UptimeFake % | LatencyToken usage | Status |
|---|---|---|---|---|
PackyCodeBiz cc | 60% | 100%Barely | 2.8sAverage | |
RunAPIBiz coding | 6.1-3.1% | 100%Barely | 3sAverage | |
SudoCodeBiz default | 60% | 100%Barely | 3.1sAverage | |
9527codeBiz claude | 4.50% | 100%Barely | 4.8sHigher | |
灵算Biz cc-max-满血 | 4.8+6.7% | 100%Barely | 3.5sMuch higher | |
UU APIBiz CC满血MAX -限cc内部 | 5.4-10% | 100%Barely | 3.3sAverage |
| Site | Channel Name | Price | Trend | Token usage | Uptime | Fake % | Latency | Status |
|---|---|---|---|---|---|---|---|---|
RunAPIBiz | codex | 1.7 | -3.1% | Much lower | 100% | Barely | 2.6s | |
UU APIBiz | Codex-GPT Pro号池 | 1 | +33% | Lower | 98% | Barely | 3.9s | |
WorldBase.aiBiz | openai | 1.7 | 0% | Much lower | 100% | Barely | 2.1s | |
球球TokenBiz | CodeX专用-Pro号池-稳定 | 0.95 | 0% | Much lower | 100% | Barely | 3s | |
luchikeyBiz | GPT-Pro通道 | 0.72 | 0% | Much lower | 98% | Barely | 4.7s | |
ModConBiz | codex-pro-官池 | 0.7 | +75% | Higher | 100% | Barely | 3.8s |
SiteChannel Name | PriceTrend | UptimeFake % | LatencyToken usage | Status |
|---|---|---|---|---|
RunAPIBiz codex | 1.7-3.1% | 100%Barely | 2.6sMuch lower | |
UU APIBiz Codex-GPT Pro号池 | 1+33% | 98%Barely | 3.9sLower | |
WorldBase.aiBiz openai | 1.70% | 100%Barely | 2.1sMuch lower | |
球球TokenBiz CodeX专用-Pro号池-稳定 | 0.950% | 100%Barely | 3sMuch lower | |
luchikeyBiz GPT-Pro通道 | 0.720% | 98%Barely | 4.7sMuch lower | |
ModConBiz codex-pro-官池 | 0.7+75% | 100%Barely | 3.8sHigher |
量子位: HuggingFace CEO力荐，Bengio团队也押注：这个1500美元训出的HRM模型，凭什么火了？
量子位: Agent终于长出了身体：Jiuwen Symbiosis背后的思考与实践
量子位: 2026智源大会开幕：推动AI、物理世界和生命科学“三体互动”
量子位: 马斯克远程敲钟穿了老黄的皮衣！SpaceX员工集体穿上绿鞋
如果我们把过去几年大模型的发展放在一个更长的时间周期来看，会发现它正在完成一次从数字世界向物理世界的迁徙。 在 2022 年底聊天机器人刚进入大众视野时，人机交互还局限在「你问我答」的回合制模式。到了 2024 至 2025 年，代理式 AI 迅速普及，人工智能开始具备自我驱动和自主规划任务的能力。 而站在 2026 年的当下，科技界普遍意识到， 物理 AI 的时代已经到来 。
SpaceX 上市首日收涨 19%，市值升至 2.1 万亿美元，马斯克成全球首位万亿富豪 大疆就云台相机专利起诉影石，影石反诉 余承东：要带盘古大模型从中国第一走向世界第一 胖东来：公司从未有任何降薪决定 OpenAI 允许 Codex 用户「囤」重置机会：额度耗尽时自己决定何时刷新 MiMo 负责人罗福莉：MiMo Code 由 5 人用 14 天 vibe-coding 完成 Genspark 获 1 亿美元 B 轮追加融资，估值...
我们和 vivo 以及一批 AI 行动派用户聊了聊，无关宏大叙事，而是日常工作和生活里那些具体的「小事」：贴发票、写日报，手机也可以是调度各路 Agent 的入口。 AI 不必急着颠覆世界。先把这些琐碎、真实的事办好就已经很有意义。而手机，是离这些事最近的载体。 于是折叠屏重新变得有意思——它在把一件件正在发生的「小事」，同时展开。 爱范儿视频 来，看个视频，放松下。
量子位: 神了，世界杯第一天真按千问剧本踢了
This page helps you quickly judge whether an AI gateway is forwarding a real upstream model reliably or whether it shows signs of wrapping, impersonation, protocol mismatch, or watered-down capability. We send controlled probe requests and compare response structure, knowledge behavior, identity consistency, thinking traces, and signature fingerprints. It is not a formal audit and it does not guarantee perfect accuracy, but it gives you a low-cost technical sanity check before you buy, integrate, or rely on an AI gateway for production traffic.
I have tried many AI gateways recently and recorded my usage notes in this document. You can refer to the API review document
An AI gateway sits between you and an upstream model provider. It forwards requests, normalizes billing, switches providers, or wraps multiple model backends behind one interface. That can be convenient, but it also means you are no longer talking directly to the original model vendor.
Common risks include protocol mismatches, identity substitution, abnormal knowledge behavior, modified responses, unclear logging or key handling, and uptime that depends on a hidden upstream. Some services also market one model name while routing traffic to a weaker or different backend.
We send structured probe requests and score the results across protocol consistency, response structure, knowledge answers, identity consistency, thinking traces, and signature fingerprints. Instead of trusting a single field, the tool cross-checks multiple technical signals to surface suspicious AI gateway behavior faster.
Technical metrics are not a guarantee of operator credibility. Our algorithm can identify whether an endpoint appears genuine and stable through repeated probes, but it cannot predict how long the operator will remain in business. Third-party relay services are heavily affected by upstream policy changes, and providers change frequently in this industry. For your financial safety, stay risk-aware and start with the smallest possible top-up. Only consider small renewals after you have confirmed stable long-term service, and avoid stocking up a large balance at once in case the provider shuts down unexpectedly.