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

2026/07/02 运行日期

claude-opus-4-8 官方基准

2 通道分数 < 70 与基线相似度

已揭晓揭晓进度

步骤 / 指标

ClaudeZZ

R-kiro

sucui

O-Max-NL

PackyCode

cc-Zeus

HongMaCC

u-cc

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 v2.1.195 claude-cli v2.1.195 claude-cli v2.1.195 claude-cli v2.1.195 claude-cli v2.1.195

偏差标红基于官方基准

步骤 1: ping text

时间 22:30:29 ~ 22:30:39 (5.4s) 22:30:33 ~ 22:30:38 (4.1s) 22:30:27 ~ 22:30:32 (3.0s) 22:30:31 ~ 22:30:36 (4.0s) 22:30:35 ~ 22:30:40 (3.4s)

提示词

查看本步 Prompt

ping. Reply with the single word: pong

HTTP 头(TTFB) 3,882 ms 2,838 ms 1,894 ms 2,666 ms 1,846 ms

首文本(SSE delta) 3,883 ms 3,064 ms 1,895 ms 2,667 ms 1,847 ms

总耗时 4,394 ms 3,064 ms 1,981 ms 2,667 ms 2,299 ms

输入 2 2 2,326 0 2,326

输出 4 4 4 22 4

缓存创建

16,029

5m: 0 / 1h: 16,029

15,319

5m: 0 / 1h: 15,319

1,711

5m: 0 / 1h: 1,711

5m: 0 / 1h: 0

缓存读取 0 0 12,732 0 14,443

响应

pong

pong

pong

pong

pong

步骤 2: identity text

时间 22:34:32 ~ 22:34:42 (8.9s) 22:32:21 ~ 22:32:31 (9.5s) 22:32:13 ~ 22:32:23 (5.1s) 22:34:07 ~ 22:34:12 (4.2s) 22:32:51 ~ 22:32:56 (3.3s)

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。严格按以下 3 行格式输出，全部小写英文，不要任何其他文字、解释、标点：
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) 7,694 ms 7,199 ms 2,074 ms 2,892 ms 1,832 ms

首文本(SSE delta) 7,695 ms 7,234 ms 2,074 ms 2,893 ms 1,833 ms

总耗时 8,054 ms 8,448 ms 3,013 ms 2,893 ms 2,231 ms

输入 2 2 2 0 2

输出 69 28 28 91 28

缓存创建

203

5m: 0 / 1h: 203

191

5m: 0 / 1h: 191

2,515

5m: 0 / 1h: 2,515

5m: 0 / 1h: 0

缓存读取 16,029 15,319 14,443 0 16,958

输出速度 - - - burst -

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

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

步骤 3: cutoff text

时间 22:37:57 ~ 22:38:07 (7.0s) 22:34:12 ~ 22:34:22 (7.2s) 22:35:05 ~ 22:35:10 (4.0s) 22:35:31 ~ 22:35:36 (4.8s) 22:36:08 ~ 22:36:13 (3.1s)

提示词

查看本步 Prompt

Knowledge cutoff date? Reply with YYYY-MM only.

HTTP 头(TTFB) 5,486 ms 5,415 ms 1,766 ms 3,727 ms 2,043 ms

首文本(SSE delta) 5,490 ms 5,440 ms 1,767 ms 3,728 ms 2,044 ms

总耗时 6,073 ms 5,817 ms 2,818 ms 3,728 ms 2,107 ms

输入 2 2 2 0 2

输出 6 6 6 33 6

缓存创建

130

5m: 0 / 1h: 130

5m: 0 / 1h: 54

5m: 0 / 1h: 0

5m: 0 / 1h: 64

缓存读取 16,232 15,510 16,958 0 16,958

输出速度 - - - burst -

响应

2025-03

2026-01

2026-01

2026-01

2026-01

步骤 4: identity_free text

时间 22:42:02 ~ 22:42:12 (7.1s) 22:37:33 ~ 22:37:43 (5.0s) 22:36:58 ~ 22:37:08 (5.0s) 22:38:43 ~ 22:38:48 (3.4s) 22:39:54 ~ 22:40:04 (5.2s)

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) 5,425 ms 2,775 ms 1,875 ms 2,225 ms 3,142 ms

首文本(SSE delta) 5,426 ms 2,800 ms 1,876 ms 2,226 ms 3,151 ms

总耗时 6,317 ms 3,904 ms 3,189 ms 2,226 ms 4,255 ms

输入 2 2 2 0 2

输出 60 66 73 62 68

缓存创建

5m: 0 / 1h: 54

5m: 0 / 1h: 39

5m: 0 / 1h: 0

5m: 0 / 1h: 35

缓存读取 16,362 15,564 17,012 0 17,022

输出速度 67.3 tok/s 59.8 tok/s 55.6 tok/s burst 61.6 tok/s

响应

我是 Claude Code，Anthropic 公司开发的官方命令行编程助手，基于 Claude 模型，帮助你完成软件工程和代码相关任务。

我是 Claude Code，Anthropic 公司开发的官方命令行 AI 编程助手，基于 Claude Opus 4.8 模型，能帮你完成软件开发任务。

你好！我是 Claude，一款由 Anthropic 开发的 AI 助手。我可以帮你编写代码、解答问题、处理文档和完成各类软件工程任务。很高兴为你服务！

我是 Claude，由 Anthropic 开发的 AI 助手。

我是 Claude，一个由 Anthropic 打造的 AI 助手，当前模型为 Opus 4.8，可通过 Claude Code 等工具协助完成软件工程等任务。

步骤 5: knowledge_recall text

时间 22:45:49 ~ 22:45:59 (7.7s) 22:39:16 ~ 22:39:31 (10.3s) 22:41:07 ~ 22:41:22 (10.9s) 22:40:38 ~ 22:40:53 (14.9s) 22:41:52 ~ 22:42:02 (9.0s)

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) 4,064 ms 2,691 ms 2,557 ms 13,585 ms 2,127 ms

首文本(SSE delta) 4,065 ms 4,123 ms 3,483 ms 13,586 ms 3,113 ms

总耗时 6,703 ms 9,296 ms 8,125 ms 13,748 ms 7,986 ms

输入 2 2 2 0 2

输出 355 692 678 1,141 671

缓存创建

12,037

5m: 0 / 1h: 12,037

4,008

5m: 0 / 1h: 4,008

5,761

5m: 0 / 1h: 5,761

5m: 0 / 1h: 0

3,886

5m: 0 / 1h: 3,886

缓存读取 0 15,933 0 0 1,876

输出速度 134.6 tok/s 133.8 tok/s 146.1 tok/s burst 137.7 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 22:49:27 ~ 22:49:42 (10.9s) 22:41:57 ~ 22:42:17 (15.3s) 22:44:38 ~ 22:44:43 (3.9s) 22:43:03 ~ 22:43:38 (31.2s) 22:43:12 ~ 22:43:22 (5.5s)

提示词

查看本步 Prompt

Count digits exactly in s. Return {"n":len(s),"c":[count0,count7,count9]}. s=288839754217721921130568691860770697987863093023783203856683559377851388

HTTP 头(TTFB) 3,403 ms 1,618 ms 1,906 ms 29,945 ms 3,945 ms

首文本(SSE delta) 8,660 ms 2,619 ms 1,907 ms 29,946 ms 3,946 ms

总耗时 9,186 ms 9,399 ms 2,870 ms 30,120 ms 4,457 ms

输入 2 2 211 0 211

输出 613 772 19 1,300 19

缓存创建

293

5m: 0 / 1h: 293

203

5m: 0 / 1h: 203

5m: 0 / 1h: 0

缓存读取 7,863 18,549 0 0 0

输出速度 burst 113.9 tok/s - burst -

响应

{"n":72,"c":[6,10,7]}

{"n":73,"c":[6,10,7]}

{"n":70,"c":[3,7,7]}

{"n":72,"c":[6,10,7]}

{"n":70,"c":[6,9,4]}

总计

总输入(含缓存) 85,244 100,701 73,770 0 73,787

总输出 1,107 1,568 808 2,649 796

总缓存创建

28,746

5m: 0 / 1h: 28,746

19,814

5m: 0 / 1h: 19,814

10,080

5m: 0 / 1h: 10,080

5m: 0 / 1h: 0

3,985

5m: 0 / 1h: 3,985

总缓存读取 56,486 80,875 61,145 0 67,257

输出速度 117.6 tok/s 117.2 tok/s 126.1 tok/s burst 123.6 tok/s

总执行时间 46.9s 51.6s 31.9s 1m 2.5s 29.4s

总墙钟时间

19m 13.1s

含步骤间等待

11m 44.0s

含步骤间等待

14m 15.9s

含步骤间等待

13m 7.0s

含步骤间等待

12m 46.9s

含步骤间等待

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.2356

Anthropic: Claude Opus 4.8

$0.2035

Anthropic: Claude Opus 4.8

$0.1265

Anthropic: Claude Opus 4.8

>=$0.0662

Anthropic: Claude Opus 4.8 (成本不完整)

$0.0912

Anthropic: Claude Opus 4.8

通道指纹

来源步骤 digit_count digit_count digit_count digit_count digit_count

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

CDN - cloudflare cloudflare - -

ID 格式 msg_01SAMmQRbn8U9nhF7YDg7KY3 msg_01DGZwbPfAo9EXozySkTyMEd msg_01U9XcRVxcLA3Zjp4EDbMMAN msg_767a15279e2f712c83fc13de msg_01PEvRRmUYKUvYd3gAp6bfaV

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-kiro	server: nginx	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	—
O-Max-NL	via: 1.1 Caddy cf-ray: a14e6a332e439131-FRA server: cloudflare x-new-api-version: v0.12.10 x-oneapi-request-id: 202607021442034673850368268d9d6HxZRJM2f	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-new-api-versionhdr: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	a14e6a332e439131-FRA
cc-Zeus	via: 1.1 Caddy cf-ray: a14e6e031c6e68e6-LAX server: cloudflare x-oneapi-request-id: 20260702224439549671783PhiUdatX	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	a14e6e031c6e68e6-LAX
u-cc	server: nginx x-new-api-version: 1.0.17 x-oneapi-request-id: 202607021443045296684148268d9d6p1gvTk4Q	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-new-api-versionhdr:x-oneapi-request-idbody:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	—
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_011CcdQdE17vKe1AX1U6gMfK

协议指纹评分（机器自动）

总分 / 100 74 87 99 54 100

基线状态已对照基线已对照基线已对照基线已对照基线已对照基线

缓存命中比 7.2 /10 ×20 10.0 /10 ×20 10.0 /10 ×20 0.0 /10 ×20 10.0 /10 ×20

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

延迟基线 4.0 /10 ×15 5.9 /10 ×15 10.0 /10 ×15 9.9 /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 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 10.0 /10 ×14

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

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

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

身份(结构) 10.0 /10 ×7 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 10.0 /10 ×7

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

知识截止 0.0 /10 ×5 10.0 /10 ×5 10.0 /10 ×5 10.0 /10 ×5 10.0 /10 ×5

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

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

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

Req-ID 透传 0.0 /10 ×2 0.0 /10 ×2 0.0 /10 ×2 0.0 /10 ×2 10.0 /10 ×2

SDK 一致 10.0 /10 ×2 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 74 v3.26.0 87 v3.26.0 99 v3.26.0 54 v3.26.0 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 Thinking 体量 thinking_volume_match
14 缓存连续 cache_continuity_intra
15 知识召回 knowledge_recall_match
16 延迟基线 latency_baseline_match
17 原生 msg-ID anthropic_msg_id_format
18 Sliding 缓存 cache_sliding_correctness
19 结构化身份 identity_structured_match
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 6788 ms 6655 ms 3666 ms 9230 ms 3889 ms

输出速度 117.6 tok/s 117.2 tok/s 126.1 tok/s burst 123.6 tok/s

完成步骤 6/6 成功 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 ⓘ

$0.2356

$0.2035

$0.1265

>=$0.0662 (成本不完整)

$0.0912

缓存读取占比 ⓘ 66% 80% 83% - 91%

揭晓

通道

服务商 ClaudeZZ ↗

通道R-kiro

服务商 sucui ↗

通道O-Max-NL

服务商 PackyCode ↗

通道cc-Zeus

服务商 HongMaCC ↗

通道u-cc

官方基准

返回列表