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

2026/06/22 运行日期

claude-opus-4-8 官方基准

2 通道分数 < 70 另含 1 任务失败

已揭晓揭晓进度

步骤 / 指标

官方基准

WorldBase.ai

M-

LingxiCode

M-

SSSAiCode

O-Max-L2

SSSAiCode

R-Kiro

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 00:59:16 ~ 00:59:27 (5.2s) 00:59:08 ~ 00:59:19 (5.7s) 00:59:14 ~ 00:59:55 (38.7s) 00:59:10 ~ 00:59:31 (15.8s) 00:59:12 ~ 00:59:18 (4.4s)

提示词

查看本步 Prompt

ping. Reply with the single word: pong

HTTP 头(TTFB) 3,546 ms 3,311 ms 32,493 ms 11,721 ms 3,048 ms

首文本(SSE delta) 3,548 ms 3,314 ms 37,187 ms 11,809 ms 3,101 ms

总耗时 3,859 ms 4,214 ms 37,416 ms 11,881 ms 3,194 ms

输入 1,454 9 1,454 1,428 0

输出 4 2 4 4 4

缓存创建

30,222

5m: 0 / 1h: 30,222

23,550

5m: 23,550 / 1h: 0

30,222

5m: 0 / 1h: 30,222

29,763

5m: 0 / 1h: 29,763

缓存读取 0 0 0 0 0

响应

pong

pong

pong

pong

pong

步骤 2: identity text

时间 01:00:35 ~ 01:00:40 (3.9s) 01:03:10 ~ 01:03:20 (7.5s) 01:03:49 ~ 01:03:59 (4.9s) 01:02:01 ~ 01:02:06 (4.9s) 01:01:59 ~ 01:02:04 (4.8s)

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。请按下列 3 行格式精确填空作答（label 不变，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) 2,566 ms 3,682 ms 3,041 ms 3,431 ms 3,532 ms

首文本(SSE delta) 2,568 ms 3,683 ms 3,058 ms 3,751 ms 3,538 ms

总耗时 2,908 ms 4,581 ms 3,936 ms 3,898 ms 3,881 ms

输入 2 11 2 2 0

输出 28 64 28 30 26

缓存创建

1,634

5m: 0 / 1h: 1,634

23,744

5m: 23,744 / 1h: 0

1,634

5m: 0 / 1h: 1,634

1,608

5m: 0 / 1h: 1,608

缓存读取 30,222 0 30,222 29,763 0

输出速度 - 71.3 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 01:03:14 ~ 01:03:19 (3.7s) 01:05:22 ~ 01:05:32 (5.5s) 01:06:09 ~ 01:06:19 (5.1s) 01:03:12 ~ 01:03:22 (5.3s) 01:05:09 ~ 01:05:14 (2.6s)

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) 2,354 ms 3,691 ms 3,997 ms 3,510 ms 1,799 ms

首文本(SSE delta) 2,356 ms 3,692 ms 3,998 ms 3,708 ms 1,850 ms

总耗时 2,569 ms 4,246 ms 4,257 ms 3,770 ms 1,854 ms

输入 2 11 2 2 0

输出 6 11 6 6 8

缓存创建

5m: 0 / 1h: 62

24,142

5m: 24,142 / 1h: 0

5m: 0 / 1h: 62

5m: 0 / 1h: 64

缓存读取 31,856 0 31,856 31,371 0

响应

2026-01

2026-01

2026-01

2025-01

2026-01

步骤 4: identity_free text

时间 01:05:03 ~ 01:05:13 (5.0s) 01:07:16 ~ 01:07:26 (6.5s) 01:09:49 ~ 01:09:59 (5.9s) 01:04:32 ~ 01:04:37 (3.6s) 01:08:59 ~ 01:09:04 (4.3s)

提示词

查看本步 Prompt

用中文说说你的身份：什么产品？由谁打造？50 字以内。

HTTP 头(TTFB) 2,599 ms 3,266 ms 3,705 ms 2,056 ms 3,363 ms

首文本(SSE delta) 2,601 ms 3,268 ms 3,706 ms 2,698 ms 3,364 ms

总耗时 3,887 ms 5,127 ms 4,841 ms 2,764 ms 3,508 ms

输入 2 57 2 2 0

输出 57 79 70 51 22

缓存创建

5m: 0 / 1h: 35

312

5m: 312 / 1h: 0

5m: 0 / 1h: 31

缓存读取 31,918 24,142 31,918 31,435 0

输出速度 44.3 tok/s 42.5 tok/s 61.7 tok/s - -

响应

我是 Claude，由 Anthropic 打造的 AI 助手，当前模型为 Claude Opus 4.8，可用于编程及多种任务协助。

我是 Kiro，一款由 Anthropic 开发的AI 编程助手，基于 Claude 模型构建，帮助开发者完成代码编写、分析、规划等软件工程任务。

我是 Claude，一个由 Anthropic 开发的 AI 助手，具体型号为 Opus 4.8（claude-opus-4-8）。我能协助完成软件工程及多种任务。

我是 Claude，一个由 Anthropic 开发的人工智能助手。我可以帮助处理写作、编程、分析等各种任务。

我是 Claude,一个由 Anthropic 开发的 AI 助手。

步骤 5: knowledge_recall text

时间 01:07:07 ~ 01:07:17 (8.9s) 01:09:04 ~ 01:09:39 (32.0s) 01:11:17 ~ 01:11:27 (8.7s) 01:06:25 ~ 01:06:35 (8.2s) 01:11:56 ~ 01:12:06 (8.1s)

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) 2,782 ms 3,369 ms 2,728 ms 5,067 ms 4,403 ms

首文本(SSE delta) 5,591 ms 3,375 ms 5,602 ms 5,259 ms 6,460 ms

总耗时 7,888 ms 30,958 ms 7,856 ms 7,362 ms 7,336 ms

输入 2 10 2 2 0

输出 672 2,593 688 353 335

缓存创建

4,764

5m: 0 / 1h: 4,764

4,736

5m: 4,736 / 1h: 0

4,773

5m: 0 / 1h: 4,773

4,271

5m: 0 / 1h: 4,271

缓存读取 0 0 0 0 0

输出速度 burst 94.0 tok/s burst burst 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

{"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 01:10:00 ~ 01:10:05 (3.0s) 01:10:51 ~ 01:11:01 (8.9s) 01:13:02 ~ 01:13:08 (3.1s) 01:09:17 ~ 01:09:22 (1.2s) 01:13:52 ~ 01:14:07 (12.2s)

提示词

查看本步 Prompt

Count digits exactly in s. Return {"n":len(s),"c":[count0,count7,count9]}. s=218533593957313083394928921447120532358852512660489495786374442767665748

HTTP 头(TTFB) 1,642 ms 5,077 ms 1,889 ms - 4,961 ms

首文本(SSE delta) 1,643 ms 5,078 ms 1,890 ms - 11,319 ms

总耗时 2,016 ms 7,833 ms 2,305 ms - 11,467 ms

输入 211 473 211 - 0

输出 19 48 19 - 20

缓存创建

5m: 0 / 1h: 0

缓存读取 0 0 0 - 0

输出速度 - 17.4 tok/s - - -

响应

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

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

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

失败

评测失败

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

总计

总输入(含缓存)

132,386

101,197

132,391

129,742

仅前 6 步

总输出

786

2,797

815

444

仅前 6 步

415

总缓存创建

36,717

5m: 0 / 1h: 36,717

76,484

5m: 76,484 / 1h: 0

36,722

5m: 0 / 1h: 36,722

35,737

5m: 0 / 1h: 35,737

仅前 6 步

总缓存读取

93,996

24,142

93,996

92,569

仅前 6 步

输出速度

44.3 tok/s

84.1 tok/s

61.7 tok/s

burst

仅前 6 步

burst

总执行时间

29.7s

1m 6.0s

1m 6.5s

39.0s

仅前 6 步

36.4s

总墙钟时间

10m 48.8s

含步骤间等待

11m 52.1s

含步骤间等待

13m 53.0s

含步骤间等待

10m 12.0s

含步骤间等待

仅前 6 步

14m 54.9s

含步骤间等待

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.3045

Anthropic: Claude Opus 4.8

$0.3836

Anthropic: Claude Opus 4.8

$0.3053

Anthropic: Claude Opus 4.8

$0.2879

Anthropic: Claude Opus 4.8 (部分)

>=$0.0104

Anthropic: Claude Opus 4.8 (成本不完整)

通道指纹

来源步骤 digit_count digit_count digit_count knowledge_recall digit_count

平台 direct (100%) direct (100%) one-api (100%) direct (100%) direct (75%)

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

CDN - cloudflare cloudflare - -

ID 格式 msg_01MtwfggDAvxXCx5zc5MB2ez msg_8dca081ba5304787a523bdd793e306bb msg_01LPJaHRgaLkfd5TEc7BfkH6 msg_01Xy3QPTSno6REVG2EU2RmiQ msg_993e4b6022014b14a11c71f8111ef242

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
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_011CcGmvR9BfBj4sMmsbKhdw
M-	cf-ray: a0f4a11018552d8e-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	32b51430-a910-4006-b6d4-fe49ce9560e5 a0f4a11018552d8e-AMS
M-	cf-ray: a0f4a446daec7343-NRT server: cloudflare set-cookie: server_name_session=46db03c0040e9983379bcced7435c21b; Max-Age=86400; httponly; path=/ x-oneapi-request-id: 20260621171304416581615kqNtvHum	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:x-oneapi-request-idhdr:server:cloudflarehdr:cf-raybody:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	a0f4a446daec7343-NRT
O-Max-L2	cf-ray: a0f49a96aa442b8c-LAX server: Photon-Edge set-cookie: _cfuvid=xg5I.xHnfL7SDjHwO.mWvkJPW4bdhKpSG16qEKFBxK8-1782061586.9878824-1.0.1.1-ITOwf8vr221HIOO0NhHLAq5ixPBuYAYubvtBH.nrR_E; 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	1782061586948-xh2i0z8ovuklb req_011CcGmeaGRCbcLF4YQkJTkH a0f49a96aa442b8c-LAX
R-Kiro	server: Photon-Edge x-amzn-requestid: 6411245d-7f0f-427e-b62b-1ca03eebd97c	id_format: msg_prefix has_service_tier: false has_inference_geo: false model_has_date_suffix: false usage_has_cache_fields: true	body:id:msg_prefixbody:usage.cache_fieldsupstream:x-amzn-requestidupstream:anthropic-direct-from-body	1782062033786-paegspxmc1h11 6411245d-7f0f-427e-b62b-1ca03eebd97c

协议指纹评分（机器自动）

总分 / 100 100 63 98 53

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 0.0 /10 ×20 10.0 /10 ×20 — 0.0 /10 ×20

缓存 TTL 一致 10.0 /10 ×15 0.0 /10 ×15 10.0 /10 ×15 — 0.0 /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 — 5.0 /10 ×14

sliding 5m 10.0 /10 ×13 5.3 /10 ×13 10.0 /10 ×13 — 5.0 /10 ×13

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

系统提示纯净 10.0 /10 ×8 0.0 /10 ×8 10.0 /10 ×8 — 未计入

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 0.0 /10 ×7 10.0 /10 ×7 — 10.0 /10 ×7

service_tier 10.0 /10 ×6 10.0 /10 ×6 10.0 /10 ×6 — 0.0 /10 ×6

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

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

流式投递 10.0 /10 ×5 8.2 /10 ×5 10.0 /10 ×5 — 2.7 /10 ×5

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

与基线相似度基准（参考） 63 v3.24.0 98 v3.24.0 执行失败，不评分 53 v3.24.0

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 信封自报一致 envelope_self_report_match
20 中文遵循 instruction_following_lang
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 3855 ms 9493 ms 10102 ms 5935 ms 5207 ms

输出速度 44.3 tok/s 84.1 tok/s 61.7 tok/s burst burst

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 ⓘ

$0.3045

$0.3836

$0.3053

$0.2879 (部分)

>=$0.0104 (成本不完整)

缓存读取占比 ⓘ 71% 24% 71% 71% (部分) -

揭晓

通道

官方基准

服务商 WorldBase.ai ↗

通道M-

服务商 LingxiCode ↗

通道M-

服务商 SSSAiCode ↗

通道O-Max-L2

服务商 SSSAiCode ↗

通道R-Kiro

返回列表