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。

5 通道 · 6 步骤

2026/06/26 运行日期

claude-opus-4-8 官方基准

3 通道分数 < 70 另含 1 任务失败

已揭晓揭晓进度

步骤 / 指标

FoxCode

R-ultra

SudoCode

M-Max

WorldBase.ai

M-

SSSAiCode

O-Max-L2

官方基准

基本信息

服务名 Claude Code 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 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 claude-cli v2.1.160

偏差标红基于官方基准

步骤 1: ping text

时间 21:43:36 ~ 21:43:41 (3.5s) 21:43:38 ~ 21:43:43 (3.6s) 21:43:34 ~ 21:43:44 (5.2s) 21:43:32 ~ 21:43:37 (3.6s) 21:43:40 ~ 21:43:50 (5.0s)

提示词

查看本步 Prompt

ping. Reply with the single word: pong

HTTP 头(TTFB) 2,290 ms 2,157 ms 3,573 ms 2,669 ms 3,883 ms

首文本(SSE delta) 2,419 ms 2,159 ms 3,574 ms 2,687 ms 3,885 ms

总耗时 2,502 ms 2,301 ms 3,641 ms 2,848 ms 4,036 ms

输入 80 2 3,636 1,428 1,454

输出 2 4 2 4 4

缓存创建

23,480

5m: 23,480 / 1h: 0

31,672

5m: 0 / 1h: 31,672

20,601

5m: 20,601 / 1h: 0

29,758

5m: 0 / 1h: 29,758

30,222

5m: 0 / 1h: 30,222

缓存读取 0 0 0 0 0

响应

pong

pong

pong

pong

pong

步骤 2: identity text

时间 21:46:04 ~ 21:46:09 (4.6s) 21:44:46 ~ 21:44:51 (3.6s) 21:46:39 ~ 21:46:49 (8.4s) 21:46:43 ~ 21:46:48 (3.9s) 21:45:03 ~ 21:45:08 (3.8s)

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。自报身份，输出严格遵守以下 3 行格式（不允许 markdown、代码块、额外文本）：
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) 2,410 ms 1,962 ms 4,274 ms 2,706 ms 1,988 ms

首文本(SSE delta) 2,411 ms 1,963 ms 4,275 ms 2,985 ms 1,990 ms

总耗时 3,913 ms 2,359 ms 6,886 ms 3,119 ms 2,840 ms

输入 33 2 3,668 2 2

输出 98 28 16 30 28

缓存创建

23,734

5m: 23,734 / 1h: 0

31,866

5m: 0 / 1h: 31,866

20,785

5m: 20,785 / 1h: 0

1,620

5m: 0 / 1h: 1,620

1,646

5m: 0 / 1h: 1,646

缓存读取 0 0 0 29,758 30,222

输出速度 65.2 tok/s - - - -

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-4-20250514

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

步骤 3: cutoff text

时间 21:48:18 ~ 21:48:24 (4.4s) 21:47:42 ~ 21:47:47 (3.1s) 21:48:34 ~ 21:48:49 (10.2s) 21:50:41 ~ 21:50:46 (4.0s) 21:46:17 ~ 21:46:27 (5.0s)

提示词

查看本步 Prompt

你掌握的世界知识截止于何时？YYYY-MM 格式作答，不解释。

HTTP 头(TTFB) 2,616 ms 1,962 ms 5,401 ms 2,736 ms 3,448 ms

首文本(SSE delta) 2,617 ms 1,991 ms 6,396 ms 2,855 ms 3,450 ms

总耗时 3,636 ms 2,109 ms 7,184 ms 2,963 ms 4,008 ms

输入 79 2 3,678 2 2

输出 35 6 3 6 6

缓存创建

398

5m: 398 / 1h: 0

31,920

5m: 0 / 1h: 31,920

487

5m: 487 / 1h: 0

5m: 0 / 1h: 56

5m: 0 / 1h: 64

缓存读取 23,734 0 20,350 31,378 31,868

输出速度 34.3 tok/s - - - -

响应

2026-01

2026-01

2026-01

2025-03

2026-01

步骤 4: identity_free text

时间 21:50:10 ~ 21:50:20 (5.2s) 21:49:54 ~ 21:49:59 (4.2s) 21:50:24 ~ 21:50:34 (6.0s) 21:53:19 ~ 21:53:24 (4.6s) 21:50:19 ~ 21:50:24 (3.7s)

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) 2,754 ms 2,078 ms 3,530 ms 2,837 ms 1,600 ms

首文本(SSE delta) 2,755 ms 2,079 ms 3,531 ms 3,663 ms 1,601 ms

总耗时 4,408 ms 3,142 ms 3,604 ms 3,831 ms 2,788 ms

输入 79 2 3,716 2 2

输出 86 63 19 55 72

缓存创建

384

5m: 384 / 1h: 0

31,959

5m: 0 / 1h: 31,959

21,052

5m: 21,052 / 1h: 0

5m: 0 / 1h: 31

5m: 0 / 1h: 39

缓存读取 24,132 0 0 31,434 31,932

输出速度 52.0 tok/s 59.3 tok/s - - 60.7 tok/s

响应

我是 Claude，由 Anthropic 开发的 AI 助手，当前运行的模型版本为 claude-opus-4-8，致力于帮助用户完成各类专业任务。

你好！我是 Claude，由 Anthropic 开发的 AI 助手。我可以帮你处理软件工程、写作、分析等各类任务。很高兴为你服务！

我是 Claude，一个 AI 助手，由 Anthropic 公司开发。

我是 Claude，一个由 Anthropic 开发的人工智能助手。我可以帮助你处理写作、分析、编程、问答等各类任务。

我是 Claude，由 Anthropic 开发的 AI 助手。我可以帮助你完成软件工程、写作、分析等多种任务，致力于提供有帮助、诚实且安全的回应。

步骤 5: knowledge_recall text

时间 21:51:51 ~ 21:52:46 (54.2s) 21:51:44 ~ 21:51:59 (13.2s) 21:53:58 ~ 21:54:13 (10.9s) 21:54:51 ~ 21:55:01 (5.4s) 21:53:45 ~ 21:53:55 (6.8s)

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) 2,131 ms 5,764 ms 6,306 ms 1,942 ms 2,895 ms

首文本(SSE delta) 2,132 ms 10,694 ms 6,307 ms 4,571 ms 2,897 ms

总耗时 53,490 ms 12,225 ms 8,902 ms 4,722 ms 5,807 ms

输入 32 2 578 2 2

输出 2,572 694 348 317 327

缓存创建

4,793

5m: 4,793 / 1h: 0

4,776

5m: 0 / 1h: 4,776

3,274

5m: 3,274 / 1h: 0

4,274

5m: 0 / 1h: 4,274

4,797

5m: 0 / 1h: 4,797

缓存读取 0 0 0 0 0

输出速度 50.1 tok/s burst 134.1 tok/s - 112.4 tok/s

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

{"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 21:55:25 ~ 21:55:50 (20.7s) 21:55:56 ~ 21:56:01 (3.4s) 21:57:41 ~ 21:57:56 (14.7s) 21:57:57 ~ 21:58:02 (0.9s) 21:56:21 ~ 21:56:26 (2.9s)

提示词

查看本步 Prompt

Count digits exactly in s. Return {"n":len(s),"c":[count0,count7,count9]}. s=932308148371995639458634883392486186136279113051782973308954549791767354

HTTP 头(TTFB) 3,021 ms 2,132 ms 6,984 ms - 1,138 ms

首文本(SSE delta) 3,022 ms 2,134 ms 12,554 ms - 1,140 ms

总耗时 19,881 ms 2,472 ms 12,588 ms - 1,856 ms

输入 78 211 206 - 211

输出 2,053 19 7 - 19

缓存创建

132

5m: 132 / 1h: 0

5m: 0 / 1h: 0

缓存读取 0 0 0 - 0

输出速度 121.8 tok/s - - - -

响应

{"n":72,"c":[3,7,10]}

{"n":72,"c":[6,8,11]}

{"n":72,"c":[3,7,10]}

失败

评测失败

{"n":72,"c":[3,8,9]}

总计

总输入(含缓存)

101,168

132,414

102,031

129,745

仅前 6 步

132,463

总输出

4,846

814

395

412

仅前 6 步

456

总缓存创建

52,921

5m: 52,921 / 1h: 0

132,193

5m: 0 / 1h: 132,193

66,199

5m: 66,199 / 1h: 0

35,739

5m: 0 / 1h: 35,739

仅前 6 步

36,768

5m: 0 / 1h: 36,768

总缓存读取

47,866

20,350

92,570

仅前 6 步

94,022

输出速度

66.9 tok/s

59.3 tok/s

134.1 tok/s

仅前 6 步

97.4 tok/s

总执行时间

1m 32.6s

31.1s

55.5s

22.4s

仅前 6 步

27.2s

总墙钟时间

12m 14.2s

含步骤间等待

12m 22.9s

含步骤间等待

14m 22.0s

含步骤间等待

14m 29.9s

含步骤间等待

仅前 6 步

12m 45.9s

含步骤间等待

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.3537

Anthropic: Claude Opus 4.8

$0.8477

Anthropic: Claude Opus 4.8

$0.3560

Anthropic: Claude Opus 4.8

$0.2871

Anthropic: Claude Opus 4.8 (部分)

$0.2966

Anthropic: Claude Opus 4.8

通道指纹

来源步骤 digit_count digit_count digit_count knowledge_recall digit_count

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

上游 Anthropic 直连 (100%) Anthropic 直连 (75%) Anthropic 直连 (100%) Anthropic 直连 (100%) Anthropic 直连 (100%)

CDN cloudflare - cloudflare - -

ID 格式 msg_01rhguhRXF5iGB2wvWK1CYfc msg_vrtx_01Ru6UTPwgY4Vsbdm3gctpos msg_V2AFFTLr5emARpjPvW5GiYEx msg_01Qr3sHZsK5M5sY8gn5xdosb msg_013NBGpmReDxQnCCs3mABais

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
R-ultra	cf-ray: a11cb5ab2f51e393-NRT server: cloudflare x-new-api-version: v0.0.0 x-oneapi-request-id: 202606261355267613433458268d9d6QrE1EcAr	id_format: msg_prefix has_service_tier: true has_inference_geo: true model_has_date_suffix: false usage_has_cache_fields: true	hdr:x-new-api-versionhdr:x-oneapi-request-idhdr:server:cloudflarehdr:cf-raybody:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:usage.cache_fieldsupstream:anthropic-direct-from-body	a11cb5ab2f51e393-NRT
M-Max	server: nginx x-new-api-version: v0.0.0 x-oneapi-request-id: 20260626215557597072380WjjaSLXb	id_format: msg_prefix extra_fields: ["stop_details"] has_service_tier: false has_inference_geo: false model_has_date_suffix: false usage_has_cache_fields: true	hdr:x-new-api-versionhdr:x-oneapi-request-idbody:id:msg_prefixbody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	—
M-	cf-ray: a11cb904c803fe98-AMS server: cloudflare	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	hdr:server:cloudflarehdr:cf-raybody:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	fa57fdc6-470c-4d7c-a9e6-48a919fc148a a11cb904c803fe98-AMS
O-Max-L2	cf-ray: a11cb4d65c179898-LAX server: Photon-Edge set-cookie: _cfuvid=nq6gEW1OIunbnP1l062Yn5LxigFpJsha9pmLIVahboc-1782482092.5331411-1.0.1.1-3R3eFGDSXOEUxBz0oxabuQh_TM3GG0zDi_OJK2M.3_M; HttpOnly; SameSite=None; Secure; Path=/; Domain=api.anthropic.com	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	1782482092167-36q2etghcu5ml req_011CcRz5r5JsuHTEnh94CC8X a11cb4d65c179898-LAX
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_011CcRzCVYEBfRNZayoWik7a

协议指纹评分（机器自动）

总分 / 100 66 66 58 96

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 0.1 /10 ×20 0.0 /10 ×20 0.0 /10 ×20 — 8.6 /10 ×20

缓存 TTL 一致 0.0 /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 7.7 /10 ×13 0.0 /10 ×13 5.3 /10 ×13 — 10.0 /10 ×13

原生 msg-ID 10.0 /10 ×8 0.0 /10 ×8 0.0 /10 ×8 — 10.0 /10 ×8

系统提示纯净 0.0 /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 0.0 /10 ×6 10.0 /10 ×6 — 10.0 /10 ×6

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

延迟基线 10.0 /10 ×5 10.0 /10 ×5 8.3 /10 ×5 — 10.0 /10 ×5

流式投递 10.0 /10 ×5 4.9 /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 66 v3.25.0 66 v3.25.0 58 v3.25.0 执行失败，不评分基准（参考）

雷达图例

共 22 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 自报身份一致 self_identity_consistency
20 信封自报一致 envelope_self_report_match
21 档位思考量 tier_thinking_volume_match
22 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 14638 ms 4101 ms 7134 ms 3497 ms 3556 ms

输出速度 66.9 tok/s 59.3 tok/s 134.1 tok/s - 97.4 tok/s

完成步骤 6/6 成功 6/6 成功 6/6 成功 5/6 成功 · 1 失败 6/6 成功

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.3537

$0.8477

$0.3560

$0.2871 (部分)

$0.2966

缓存读取占比 ⓘ 47% 0% 20% 71% (部分) 71%

揭晓

通道

服务商 FoxCode ↗

通道R-ultra

服务商 SudoCode ↗

通道M-Max

服务商 WorldBase.ai ↗

通道M-

服务商 SSSAiCode ↗

通道O-Max-L2

官方基准

返回列表