不检测渠道接口规范, 只检测渠道任务能力, 通过多种渠道统计渠道差异
数据更新于 · 11 天前
测试次数
1,828
测试题目
20
总计花费
$1141.39
| 站点 | 渠道 | 模型 | 通过率 | 平均花费 (USD) | Token TPS | 测试次数 |
|---|---|---|---|---|---|---|
| CCTQ | OFFICIAL | gpt-5.6-sol | 100% | $0.0958 | 28.6 | 15 |
| Yes Code | OFFICIAL | gpt-5.6-sol | 93.3% | $0.3218 | 30.3 | 15 |
| Micu | OFFICIAL | gpt-5.6-sol | 86.7% | $0.0959 | 27.9 | 15 |
| Packy Code | OFFICIAL | gpt-5.6-sol | 80% | $0.3990 | 38.6 | 15 |
| 78 Code | OFFICIAL | gpt-5.6-sol | 80% | $0.1048 | 30.6 | 15 |
| Duck Code | OFFICIAL | gpt-5.6-sol | 80% | $0.4783 | 27.9 | 15 |
| Yunwu | OFFICIAL | claude-opus-5 | 80% | $0.5225 | 33.0 | 10 |
| CCTQ | OFFICIAL | claude-opus-5 | 66.7% | $1.2821 | 57.4 | 15 |
| TimiCC | OFFICIAL | gpt-5.6-sol | 66.7% | $0.2988 | 27.6 | 15 |
| SSSAI Code | OFFICIAL | gpt-5.6-sol | 66.7% | $0.1172 | 29.4 | 15 |
| ByteCat | OFFICIAL | claude-opus-5 | 66.7% | $1.0699 | 55.7 | 15 |
| ByteCat | OFFICIAL | gpt-5.6-sol | 66.7% | $0.1068 | 27.0 | 15 |
| Packy Code | OFFICIAL | claude-opus-5 | 66.7% | $1.7587 | 49.9 | 15 |
| Neko Code | OFFICIAL | gpt-5.6-sol | 66.7% | $0.1779 | 21.8 | 15 |
| Cubence | OFFICIAL | gpt-5.6-sol | 66.7% | $0.3859 | 23.7 | 15 |
| Micu | OFFICIAL | claude-opus-5 | 66.7% | $1.6081 | 57.0 | 15 |
| Right Code | OFFICIAL | claude-opus-5 | 60% | $1.5631 | 50.4 | 15 |
| APINebula | OFFICIAL | gpt-5.6-sol | 60% | $0.1954 | 32.0 | 15 |
| APINebula | OFFICIAL | claude-opus-5 | 60% | $1.6649 | 54.5 | 15 |
| 78 Code | OFFICIAL | claude-opus-5 | 60% | $0.5519 | 58.1 | 15 |
| Neko Code | OFFICIAL | claude-opus-5 | 60% | $1.2924 | 61.6 | 15 |
| IKun Code | OFFICIAL | claude-opus-5 | 60% | $3.3963 | 52.0 | 15 |
| SSSAI Code | OFFICIAL | claude-opus-5 | 53.3% | $1.1305 | 55.6 | 15 |
| TimiCC | OFFICIAL | claude-opus-5 | 53.3% | $0.8674 | 54.0 | 15 |
| Somebody | AWSQ | claude-opus-5 | 53.3% | $1.2743* | 44.2 | 15 |
| Right Code | OFFICIAL | gpt-5.6-sol | 53.3% | $0.0856 | 21.4 | 15 |
| Cubence | OFFICIAL | claude-opus-5 | 53.3% | $1.3276 | 50.5 | 15 |
| Duck Code | OFFICIAL | claude-opus-5 | 53.3% | $1.2128 | 53.1 | 15 |
| Yes Code | OFFICIAL | claude-opus-5 | 46.7% | $0.7560 | 47.9 | 15 |
| IKun Code | OFFICIAL | gpt-5.6-sol | 46.7% | $0.1885 | 24.9 | 15 |
花费为模型官方牌价 × 站点最近已知折扣的折算值；带 * 的行未匹配到站点折扣，按官方牌价估算。
按全部历史 trial 聚合的渠道排名，口径见下方评测方法论。
| # | 渠道 | 模型 | 通过率 | 速度 | 成本 |
|---|---|---|---|---|---|
| 1 | CCTQ · OFFICIAL | gpt-5.6-sol | 100.0% | 28.6 | $0.0958 |
| 2 | Yes Code · OFFICIAL | gpt-5.6-sol | 93.3% | 30.3 | $0.3218 |
| 3 | Micu · OFFICIAL | gpt-5.6-sol | 86.7% | 27.9 | $0.0959 |
| 4 | Packy Code · OFFICIAL | gpt-5.6-sol | 80.0% | 38.6 | $0.3990 |
| 5 | 78 Code · OFFICIAL | gpt-5.6-sol | 80.0% | 30.6 | $0.1048 |
| 6 | Duck Code · OFFICIAL | gpt-5.6-sol | 80.0% | 27.9 | $0.4783 |
| 7 | Yunwu · OFFICIAL | claude-opus-5 | 80.0% | 33.0 | $0.5225 |
| 8 | CCTQ · OFFICIAL | claude-opus-5 | 66.7% | 57.4 | $1.2821 |
| 9 | TimiCC · OFFICIAL | gpt-5.6-sol | 66.7% | 27.6 | $0.2988 |
| 10 | SSSAI Code · OFFICIAL | gpt-5.6-sol | 66.7% | 29.4 | $0.1172 |
| # | 渠道 | 模型 | 通过率 | 速度 | 成本 |
|---|---|---|---|---|---|
| 1 | Neko Code · OFFICIAL | claude-opus-5 | 60.0% | 61.6 | $1.2924 |
| 2 | 78 Code · OFFICIAL | claude-opus-5 | 60.0% | 58.1 | $0.5519 |
| 3 | CCTQ · OFFICIAL | claude-opus-5 | 66.7% | 57.4 | $1.2821 |
| 4 | Micu · OFFICIAL | claude-opus-5 | 66.7% | 57.0 | $1.6081 |
| 5 | ByteCat · OFFICIAL | claude-opus-5 | 66.7% | 55.7 | $1.0699 |
| 6 | SSSAI Code · OFFICIAL | claude-opus-5 | 53.3% | 55.6 | $1.1305 |
| 7 | APINebula · OFFICIAL | claude-opus-5 | 60.0% | 54.5 | $1.6649 |
| 8 | TimiCC · OFFICIAL | claude-opus-5 | 53.3% | 54.0 | $0.8674 |
| 9 | Duck Code · OFFICIAL | claude-opus-5 | 53.3% | 53.1 | $1.2128 |
| 10 | IKun Code · OFFICIAL | claude-opus-5 | 60.0% | 52.0 | $3.3963 |
| # | 渠道 | 模型 | 通过率 | 速度 | 成本 |
|---|---|---|---|---|---|
| 1 | Right Code · OFFICIAL | gpt-5.6-sol | 53.3% | 21.4 | $0.0856 |
| 2 | CCTQ · OFFICIAL | gpt-5.6-sol | 100.0% | 28.6 | $0.0958 |
| 3 | Micu · OFFICIAL | gpt-5.6-sol | 86.7% | 27.9 | $0.0959 |
| 4 | 78 Code · OFFICIAL | gpt-5.6-sol | 80.0% | 30.6 | $0.1048 |
| 5 | ByteCat · OFFICIAL | gpt-5.6-sol | 66.7% | 27.0 | $0.1068 |
| 6 | SSSAI Code · OFFICIAL | gpt-5.6-sol | 66.7% | 29.4 | $0.1172 |
| 7 | Neko Code · OFFICIAL | gpt-5.6-sol | 66.7% | 21.8 | $0.1779 |
| 8 | IKun Code · OFFICIAL | gpt-5.6-sol | 46.7% | 24.9 | $0.1885 |
| 9 | APINebula · OFFICIAL | gpt-5.6-sol | 60.0% | 32.0 | $0.1954 |
| 10 | TimiCC · OFFICIAL | gpt-5.6-sol | 66.7% | 27.6 | $0.2988 |
截至 2026年8月8日，本站已对 30 个渠道配置在 20 道真实编程任务上累计执行 1828 次测试，总花费约 $1141，全部自费、无中转赞助。
以测试题为主键，展示在 terminal-bench / aider-polyglot 等真实编程任务下的跑分，而非接口连通性检测。
每次 trial 在隔离容器内运行：CLI Agent（codex / claude-code 等）通过被测渠道调用模型完成任务，随后由独立 Verifier 执行任务自带的验收脚本判分，全程记录 Token 消耗、耗时与实际花费。
仅 Verifier 验收通过的 trial 计为通过；模型未完成任务、运行异常或超时一律计为失败，不做人工补测，也不挑选成绩重跑。
历史通过率 = 通过 trial 数 ÷ 总 trial 数 × 100%；水平线 = 所有 config 历史通过率的算术平均。
统计每个渠道（config）的全部历史 trial；卡片明细基于最新一次已上架 run 的切片，历史通过率跨全部已上架 run 累计。同 task/config 历史 trial 数少于 3 次标记「样本不足」，避免小样本误导。指标含通过率、平均 Token、平均花费、Token TPS。
