模型对比 | RPDiag

测试说明点击展开 5 步 quick-probe-v1 详解

前 5 步同一会话，第 6 步独立新会话，覆盖 6 个独立信号： 1. ping/pong —— 单字指令遵循 + 建立缓存上下文。 2. 身份（结构化）—— 三行 vendor / brand / model 格式，机器可解析。 3. 知识截止 —— 跨越 5 分钟 cache 边界后追问，检测 sliding 5m cache 是否真正命中。 4. 身份（自由格式）—— 自然语言自报身份，给包装层（Kiro 等）暴露品牌的机会。 5. 世界知识层级 —— 5 道公共事件硬事实题，按答对档位映射实测 tier，与请求模型对照。 6. 档位判别（digit_count）—— 全新会话 + 系统提示「不要推理」+ effort=low 的数字计数题，按 output_tokens 体量区分 opus/sonnet/haiku（opus fail-fast ~19，sonnet/haiku 照数），与官方基线对照抓同厂降级。前 5 步之间随机延迟 1–4 分钟。第 3 步（cutoff）累计跨度 > 5 分钟保住 sliding cache 检测；总跨度 > 6 分钟。第 6 步是独立新会话（不复用前 5 轮上下文，否则 opus 不再 fail-fast）。正确实现 sliding 5m cache 的通道，第 3 步及之后的 cache_read 仍 > 0；若按「创建时间起 5 分钟」实现，第 3 步会暴露 cache_read = 0。

4 通道 · 6 步骤

2026/06/27 运行日期

claude-opus-4-8 官方基准

1 通道分数 < 70 另含 1 任务失败

已揭晓揭晓进度

⚠ 1 通道不可测 · 已从主对比表移除点击展开原因

TopRouterCN / O-Max
```
评测失败
```

步骤 / 指标

0-0

O-Max

PackyCode

cc-Zeus

luminai

M-Kiro/Antigravity

官方基准

基本信息

服务名 Claude Code Claude Code Claude Code Claude Code

模型 claude-opus-4-8 claude-opus-4-8 claude-opus-4-8 claude-opus-4-8

客户端 claude-cli v2.1.160 claude-cli v2.1.160 claude-cli v2.1.160 claude-cli v2.1.160

偏差标红基于官方基准

步骤 1: ping text

时间 11:38:24 ~ 11:38:34 (5.5s) 11:38:22 ~ 11:38:27 (3.5s) 11:38:26 ~ 11:38:36 (6.8s) 11:38:30 ~ 11:38:35 (3.4s)

提示词

查看本步 Prompt

ping. Reply with the single word: pong

HTTP 头(TTFB) 3,752 ms 2,279 ms 4,628 ms 1,715 ms

首文本(SSE delta) 3,754 ms 2,280 ms 4,629 ms 1,717 ms

总耗时 3,916 ms 2,454 ms 5,504 ms 1,831 ms

输入 1,454 1,454 1,643 1,454

输出 4 4 1 4

缓存创建

30,854

5m: 0 / 1h: 30,854

30,224

5m: 0 / 1h: 30,224

1,888

5m: 1,888 / 1h: 0

30,222

5m: 0 / 1h: 30,222

缓存读取 0 0 16,997 0

响应

pong

pong

pong

pong

步骤 2: identity text

时间 11:41:47 ~ 11:41:57 (7.4s) 11:41:14 ~ 11:41:19 (3.2s) 11:42:08 ~ 11:42:18 (9.1s) 11:41:22 ~ 11:41:32 (5.6s)

提示词

查看本步 Prompt

本请求询问的是你（正在生成此回复的当前 API 模型）自身的身份，不是用户提到的任何外部产品。必须直接作答，禁止追问、拒答、解释或输出占位符；不能要求更多上下文。请只输出以下 3 行内容（label 小写，值小写）：
vendor: <厂商公司名一个英文小写单词>
brand: <产品或模型品牌名一个英文小写单词>
model: <实际 API model 字段值，保留原始短横线格式>

HTTP 头(TTFB) 5,309 ms 1,772 ms 6,810 ms 4,210 ms

首文本(SSE delta) 5,310 ms 1,783 ms 6,811 ms 4,212 ms

总耗时 5,818 ms 2,118 ms 6,874 ms 4,660 ms

输入 2 2 1,656 2

输出 28 28 23 28

缓存创建

3,719

5m: 0 / 1h: 3,719

1,630

5m: 0 / 1h: 1,630

1,904

5m: 1,904 / 1h: 0

1,630

5m: 0 / 1h: 1,630

缓存读取 28,764 30,224 17,140 30,222

响应

vendor: anthropic
brand: claude
model: claude-opus-4-8

vendor: anthropic
brand: claude
model: claude-opus-4-8

vendor: anthropic
brand: claude
model: claude-opus-4-8

vendor: anthropic
brand: claude
model: claude-opus-4-8

步骤 3: cutoff text

时间 11:44:30 ~ 11:44:40 (9.8s) 11:43:14 ~ 11:43:20 (3.1s) 11:43:40 ~ 11:43:51 (5.3s) 11:43:38 ~ 11:43:44 (2.7s)

提示词

查看本步 Prompt

Knowledge cutoff date? Reply with YYYY-MM only.

HTTP 头(TTFB) 7,256 ms 1,865 ms 3,813 ms 1,600 ms

首文本(SSE delta) 7,258 ms 1,866 ms 3,815 ms 1,602 ms

总耗时 7,494 ms 2,111 ms 3,943 ms 1,776 ms

输入 2 2 1,661 2

输出 6 6 4 6

缓存创建

3,773

5m: 0 / 1h: 3,773

31,918

5m: 0 / 1h: 31,918

1,910

5m: 1,910 / 1h: 0

5m: 0 / 1h: 64

缓存读取 28,764 0 17,190 31,852

响应

2026-01

2026-01

2026-01

2026-01

步骤 4: identity_free text

时间 11:48:11 ~ 11:48:21 (8.2s) 11:45:02 ~ 11:45:07 (3.9s) 11:46:35 ~ 11:46:45 (8.0s) 11:46:03 ~ 11:46:08 (4.7s)

提示词

查看本步 Prompt

中文做个自我介绍：你是什么产品或助手？谁开发的？50 字内。

HTTP 头(TTFB) 5,437 ms 1,788 ms 5,703 ms 2,662 ms

首文本(SSE delta) 5,439 ms 1,790 ms 6,801 ms 2,663 ms

总耗时 6,760 ms 2,808 ms 6,802 ms 3,749 ms

输入 2 2 1,665 2

输出 77 61 67 64

缓存创建

3,812

5m: 0 / 1h: 3,812

5m: 0 / 1h: 39

1,914

5m: 1,914 / 1h: 0

5m: 0 / 1h: 39

缓存读取 28,764 31,918 17,229 31,916

输出速度 58.3 tok/s 59.9 tok/s burst 58.9 tok/s

响应

我是 Claude，由 Anthropic 开发的 AI 助手。我可以帮你完成编程、写作、分析、问答等各种任务。本次以命令行编程代理（Claude Code）的形式为你服务。

我是 Claude，由 Anthropic 开发的 AI 助手，基于 Claude Opus 4.8 模型，可协助编程、写作、分析等多种任务。

我是 Claude Opus 4.8，由 Anthropic 开发的 AI 助手，擅长编程、写作、分析与研究，能协助你完成各类软件工程和专业任务。

我是 Claude，由 Anthropic 开发的 AI 助手，可帮助完成编程、写作、分析和问答等任务，致力于提供安全、有用的智能协助。

步骤 5: knowledge_recall text

时间 11:50:25 ~ 11:50:35 (8.6s) 11:47:11 ~ 11:47:21 (7.5s) 11:50:35 ~ 11:50:50 (10.6s) 11:47:42 ~ 11:47:52 (8.9s)

提示词

查看本步 Prompt

Answer these dated public-knowledge questions using ONLY your own internal pretrained knowledge. Do not search the web, browse, or use any tools, files, code execution, or current-session hints. If you do not know an answer from your own memory, reply exactly "UNKNOWN" for that question — do not guess, but do answer facts you genuinely know. Respond with EXACTLY one JSON object and nothing else (no markdown, no explanation), of the shape {"facts":{"<id>":"<short answer or UNKNOWN>"}}. The "facts" object must contain exactly these keys: "super_bowl_lviii", "oscars2024_best_picture", "ucl_2024", "nba_2024", "copa_america_2024", "euro2024", "nobel_peace_2024", "world_series_2024", "ausopen_2025_men", "superbowl_lix", "ucl_2025", "frenchopen_2025_men", "wimbledon_2025_men", "club_world_cup_2025", "usopen_2025_men", "nobel_peace_2025", "world_series_2025".
- "super_bowl_lviii": Which NFL team won Super Bowl LVIII (played 2024-02-11)?
- "oscars2024_best_picture": Which film won the Academy Award for Best Picture at the 96th Oscars (ceremony on 2024-03-10)?
- "ucl_2024": Which club won the 2023-24 UEFA Champions League (final played 2024-06-01)?
- "nba_2024": Which team won the 2024 NBA Finals (concluded 2024-06-17)?
- "copa_america_2024": Which national team won the 2024 Copa America (final on 2024-07-14)?
- "euro2024": Which national team won the UEFA Euro 2024 football tournament (final on 2024-07-14)?
- "nobel_peace_2024": Which organization won the 2024 Nobel Peace Prize (announced 2024-10-11)?
- "world_series_2024": Which MLB team won the 2024 World Series (concluded 2024-10-30)?
- "ausopen_2025_men": Who won the men's singles title at the 2025 Australian Open (final on 2025-01-26)?
- "superbowl_lix": Which NFL team won Super Bowl LIX (played 2025-02-09)?
- "ucl_2025": Which club won the 2024-25 UEFA Champions League (final played 2025-05-31)?
- "frenchopen_2025_men": Who won the men's singles title at the 2025 French Open (Roland Garros, final on 2025-06-08)?
- "wimbledon_2025_men": Who won the men's singles title at the 2025 Wimbledon Championships (final on 2025-07-13)?
- "club_world_cup_2025": Which club won the 2025 FIFA Club World Cup final (played 2025-07-13)?
- "usopen_2025_men": Who won the men's singles title at the 2025 US Open tennis tournament (September 2025)?
- "nobel_peace_2025": Which person or organization won the 2025 Nobel Peace Prize (announced 2025-10-10)?
- "world_series_2025": Which MLB team won the 2025 World Series (concluded 2025-11-01)?

HTTP 头(TTFB) 4,313 ms 1,630 ms 5,848 ms 1,928 ms

首文本(SSE delta) 4,314 ms 4,004 ms 5,849 ms 5,688 ms

总耗时 7,109 ms 6,482 ms 9,631 ms 7,930 ms

输入 2 2 271 2

输出 331 519 385 668

缓存创建

5,357

5m: 0 / 1h: 5,357

4,772

5m: 0 / 1h: 4,772

311

5m: 311 / 1h: 0

4,773

5m: 0 / 1h: 4,773

缓存读取 0 0 2,800 0

输出速度 118.4 tok/s burst 101.8 tok/s burst

响应

{"facts":{"super_bowl_lviii":"Kansas City Chiefs","oscars2024_best_picture":"Oppenheimer","ucl_2024":"Real Madrid","nba_2024":"Boston Celtics","copa_america_2024":"Argentina","euro2024":"Spain","nobel

{"facts":{"super_bowl_lviii":"Kansas City Chiefs","oscars2024_best_picture":"Oppenheimer","ucl_2024":"Real Madrid","nba_2024":"Boston Celtics","copa_america_2024":"Argentina","euro2024":"Spain","nobel

{"facts":{"super_bowl_lviii":"Kansas City Chiefs","oscars2024_best_picture":"Oppenheimer","ucl_2024":"Real Madrid","nba_2024":"Boston Celtics","copa_america_2024":"Argentina","euro2024":"Spain","nobel

{"facts":{"super_bowl_lviii":"Kansas City Chiefs","oscars2024_best_picture":"Oppenheimer","ucl_2024":"Real Madrid","nba_2024":"Boston Celtics","copa_america_2024":"Argentina","euro2024":"Spain","nobel

步骤 6: digit_count text

时间 11:53:07 ~ 11:53:12 (3.8s) 11:50:14 ~ 11:50:19 (3.1s) 11:52:36 ~ 11:52:51 (11.8s) 11:49:49 ~ 11:49:54 (2.7s)

提示词

查看本步 Prompt

Count digits exactly in s. Return {"n":len(s),"c":[count0,count7,count9]}. s=613539163122789747631555490000443923782222482829842452713541611932211462

HTTP 头(TTFB) 2,074 ms 1,887 ms 5,801 ms 1,514 ms

首文本(SSE delta) 2,075 ms 1,888 ms 5,802 ms 1,515 ms

总耗时 2,318 ms 2,061 ms 10,674 ms 1,764 ms

输入 793 212 189 211

输出 19 19 83 19

缓存创建

5m: 0 / 1h: 0

缓存读取 0 0 - 0

输出速度 - - 17.0 tok/s -

响应

{"n":72,"c":[6,4,4]}

{"n":72,"c":[5,4,4]}

{"n":72,"c":[4,5,6]}

{"n":70,"c":[7,4,5]}

总计

总输入(含缓存) 136,062 132,399 86,368 132,391

总输出 465 637 563 789

总缓存创建

47,515

5m: 0 / 1h: 47,515

68,583

5m: 0 / 1h: 68,583

7,927

5m: 7,927 / 1h: 0

36,728

5m: 0 / 1h: 36,728

总缓存读取 86,292 62,142 71,356 93,990

输出速度 99.1 tok/s 59.9 tok/s 54.1 tok/s 58.9 tok/s

总执行时间 43.3s 24.3s 51.5s 28.1s

总墙钟时间

14m 48.0s

含步骤间等待

11m 56.9s

含步骤间等待

14m 24.9s

含步骤间等待

11m 24.0s

含步骤间等待

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.3630

Anthropic: Claude Opus 4.8

$0.4840

Anthropic: Claude Opus 4.8

$0.1161

Anthropic: Claude Opus 4.8

$0.3046

Anthropic: Claude Opus 4.8

通道指纹

来源步骤 digit_count digit_count digit_count digit_count

平台 direct (100%) one-api (100%) one-api (100%) direct (100%)

上游 Anthropic 直连 (100%) Anthropic 直连 (100%) Anthropic 直连 (100%) Anthropic 直连 (100%)

CDN - cloudflare cloudflare -

ID 格式 msg_01UqSYStga5CTph5BCWg3XwC msg_01VKStAvpFnCdXwdTLHarkHa msg_01T4oaZdqZzmrT6euGxfNdusdq msg_01CnkYaZ6y9u9XdxhFyJJYi5

展开详细指纹（响应头 / body 特征 / 完整信号列表）

通道	response_headers_notable	response_body_traits	signals	request_id_chain
O-Max	via: 1.1 Caddy server: uvicorn, nginx/1.24.0 (Ubuntu)	id_format: msg_prefix via_chain: ["1.1 Caddy"] extra_fields: ["stop_details"] has_service_tier: true has_inference_geo: true model_has_date_suffix: false usage_has_cache_fields: true	hdr:viabody:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	eb57494008a504308e5b33499ceed7db
cc-Zeus	via: 1.1 Caddy cf-ray: a1217c8b8fdcc213-LAX server: cloudflare x-oneapi-request-id: 202606271150156240253171jTk8a0n	id_format: msg_prefix via_chain: ["1.1 Caddy"] extra_fields: ["stop_details"] has_service_tier: true has_inference_geo: true model_has_date_suffix: false usage_has_cache_fields: true	hdr:x-oneapi-request-idhdr:server:cloudflarehdr:cf-rayhdr:viabody:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	a1217c8b8fdcc213-LAX
M-Kiro/Antigravity	cf-ray: a1218003ac3a3d63-AMS server: cloudflare x-oneapi-request-id: 202606270352378040662158268d9d64laFwC6y	id_format: msg_prefix extra_fields: ["container"] has_service_tier: true has_inference_geo: false model_has_date_suffix: false usage_has_cache_fields: true	hdr:x-oneapi-request-idhdr:server:cloudflarehdr:cf-raybody:id:msg_prefixbody:usage.service_tierbody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	a1218003ac3a3d63-AMS
官方基准	server: nginx/1.22.1	id_format: msg_prefix extra_fields: ["stop_details"] has_service_tier: true has_inference_geo: true model_has_date_suffix: false usage_has_cache_fields: true	body:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	req_011CcT5ko2BGDikhJ94KmuAp

协议指纹评分（机器自动）

总分 / 100 92 90 72 100

基线状态已对照基线已对照基线已对照基线已对照基线

缓存命中比 10.0 /10 ×20 5.4 /10 ×20 10.0 /10 ×20 10.0 /10 ×20

缓存 TTL 一致 10.0 /10 ×15 10.0 /10 ×15 0.0 /10 ×15 10.0 /10 ×15

模型匹配 10.0 /10 ×14 10.0 /10 ×14 10.0 /10 ×14 10.0 /10 ×14

缓存连续性 10.0 /10 ×14 10.0 /10 ×14 10.0 /10 ×14 10.0 /10 ×14

sliding 5m 10.0 /10 ×13 7.7 /10 ×13 10.0 /10 ×13 10.0 /10 ×13

原生 msg-ID 10.0 /10 ×8 10.0 /10 ×8 10.0 /10 ×8 10.0 /10 ×8

系统提示纯净 5.0 /10 ×8 10.0 /10 ×8 0.0 /10 ×8 10.0 /10 ×8

身份(结构) 10.0 /10 ×7 10.0 /10 ×7 10.0 /10 ×7 10.0 /10 ×7

知识截止 10.0 /10 ×7 10.0 /10 ×7 10.0 /10 ×7 10.0 /10 ×7

身份(自由) 10.0 /10 ×7 10.0 /10 ×7 10.0 /10 ×7 10.0 /10 ×7

service_tier 10.0 /10 ×6 10.0 /10 ×6 10.0 /10 ×6 10.0 /10 ×6

inference_geo 10.0 /10 ×5 10.0 /10 ×5 0.0 /10 ×5 10.0 /10 ×5

延迟基线 6.4 /10 ×5 10.0 /10 ×5 5.7 /10 ×5 10.0 /10 ×5

流式投递 10.0 /10 ×5 10.0 /10 ×5 0.0 /10 ×5 10.0 /10 ×5

Req-ID 透传 0.0 /10 ×4 0.0 /10 ×4 0.0 /10 ×4 10.0 /10 ×4

stop_reason 10.0 /10 ×3 10.0 /10 ×3 10.0 /10 ×3 10.0 /10 ×3

SDK 一致 10.0 /10 ×2 10.0 /10 ×2 10.0 /10 ×2 10.0 /10 ×2

综合结论

与基线相似度 92 v3.25.0 90 v3.25.0 72 v3.25.0 基准（参考）

雷达图例

共 21 个维度（顺时针，从顶部 12 点起）— 点击展开对照

1 模型匹配 model_match
2 知识截止 cutoff_match
3 SDK 一致 sdk_consistency
4 Thinking 存在 thinking_present
5 流式投递 buffer_dump_match
6 自由身份 identity_free_clean
7 停止原因 stop_reason_present
8 系统提示纯净 system_prompt_clean
9 Service Tier service_tier_present
10 缓存命中比 cache_hit_ratio_match
11 缓存 TTL 一致 cache_ttl_consistency
12 推理地区 inference_geo_present
13 缓存连续 cache_continuity_intra
14 知识召回 knowledge_recall_match
15 延迟基线 latency_baseline_match
16 原生 msg-ID anthropic_msg_id_format
17 Sliding 缓存 cache_sliding_correctness
18 结构化身份 identity_structured_match
19 信封自报一致 envelope_self_report_match
20 档位思考量 tier_thinking_volume_match
21 Req-ID 透传 anthropic_request_id_passthrough

每根轴长度 = 该维度 0-10 分（越长越接近基线）。蓝色虚线圆 = 满分基准。具体权重见方法论页。

维度雷达

平均延迟 5569 ms 3006 ms 7238 ms 3618 ms

输出速度 99.1 tok/s 59.9 tok/s 54.1 tok/s 58.9 tok/s

完成步骤 6/6 成功 6/6 成功 6/6 成功 6/6 成功

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.3630

$0.4840

$0.1161

$0.3046

缓存读取占比 ⓘ 63% 47% 83% 71%

揭晓

通道

服务商 0-0 ↗

通道O-Max

服务商 PackyCode ↗

通道cc-Zeus

服务商 luminai ↗

通道M-Kiro/Antigravity

官方基准

返回列表