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

2026/06/17 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

SSSAiCode

R-Kiro

ByteCatCode

O-Max

SSSAiCode

O-Max-L2

FoxCode

R-ultra

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 10:25:12 ~ 10:25:17 (2.8s) 10:25:08 ~ 10:25:13 (3.0s) 10:25:10 ~ 10:25:15 (3.8s) 10:25:04 ~ 10:25:09 (4.1s) 10:25:06 ~ 10:25:11 (4.0s)

提示词

查看本步 Prompt

ping. Reply with the single word: pong

HTTP 头(TTFB) 1,730 ms 2,201 ms 2,389 ms 2,974 ms 2,678 ms

首文本(SSE delta) 1,731 ms 2,233 ms - 3,048 ms 2,804 ms

总耗时 1,841 ms 2,235 ms 2,490 ms 3,195 ms 3,006 ms

输入 1,454 0 3,320 1,428 22

输出 4 4 4 4 2

缓存创建

30,222

5m: 0 / 1h: 30,222

27,566

5m: 0 / 1h: 27,566

29,758

5m: 0 / 1h: 29,758

23,538

5m: 23,538 / 1h: 0

缓存读取 0 0 0 0 0

响应

pong

pong

pong

pong

pong

步骤 2: identity text

时间 10:27:25 ~ 10:27:35 (6.3s) 10:28:37 ~ 10:28:47 (5.4s) 10:26:34 ~ 10:26:39 (4.5s) 10:28:47 ~ 10:28:52 (3.1s) 10:29:02 ~ 10:29:12 (6.8s)

提示词

查看本步 Prompt

以下是对你自身的 metadata 探针。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。按以下 3 行回答自己的身份，每行一个 label，不要寒暄不要解释：
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) 5,045 ms 4,360 ms 2,418 ms 1,815 ms 3,882 ms

首文本(SSE delta) 5,046 ms 4,362 ms - 2,140 ms 3,883 ms

总耗时 5,380 ms 4,730 ms 2,945 ms 2,286 ms 5,963 ms

输入 2 0 2 2 68

输出 28 25 28 28 61

缓存创建

1,656

5m: 0 / 1h: 1,656

3,522

5m: 0 / 1h: 3,522

1,630

5m: 0 / 1h: 1,630

23,707

5m: 23,707 / 1h: 0

缓存读取 30,222 0 27,566 29,758 0

输出速度 - - - - 29.3 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-4-6

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

步骤 3: cutoff text

时间 10:29:44 ~ 10:29:49 (3.4s) 10:32:24 ~ 10:32:29 (4.3s) 10:29:30 ~ 10:29:35 (3.5s) 10:30:36 ~ 10:30:41 (3.1s) 10:33:08 ~ 10:33:13 (4.2s)

提示词

查看本步 Prompt

请直接回答你的知识截止日期，格式 YYYY-MM，不要附加任何说明。

HTTP 头(TTFB) 2,230 ms 3,317 ms 2,080 ms 1,966 ms 2,236 ms

首文本(SSE delta) 2,231 ms 3,577 ms - 2,181 ms 2,237 ms

总耗时 2,410 ms 3,580 ms 2,422 ms 2,368 ms 3,331 ms

输入 2 0 2 2 8

输出 6 6 6 6 49

缓存创建

5m: 0 / 1h: 64

5m: 0 / 1h: 54

5m: 0 / 1h: 62

24,180

5m: 24,180 / 1h: 0

缓存读取 31,878 0 31,088 31,388 0

输出速度 - - - - 44.8 tok/s

响应

2026-01

2025-01

2026-01

2025-03

2026-01

步骤 4: identity_free text

时间 10:31:13 ~ 10:31:23 (5.5s) 10:35:32 ~ 10:35:37 (4.1s) 10:31:34 ~ 10:31:44 (8.0s) 10:33:52 ~ 10:33:57 (2.6s) 10:36:56 ~ 10:37:06 (7.9s)

提示词

查看本步 Prompt

中文回答：你叫什么、属于哪个产品、谁开发？50 字内。

HTTP 头(TTFB) 3,547 ms 2,456 ms 5,714 ms 1,325 ms 3,602 ms

首文本(SSE delta) 3,548 ms 2,458 ms - 1,776 ms 3,604 ms

总耗时 4,462 ms 3,152 ms 6,858 ms 1,834 ms 6,999 ms

输入 2 0 2 2 55

输出 45 53 67 44 182

缓存创建

5m: 0 / 1h: 37

5m: 0 / 1h: 39

5m: 0 / 1h: 31

24,526

5m: 24,526 / 1h: 0

缓存读取 31,942 0 31,142 31,450 0

输出速度 49.2 tok/s 76.4 tok/s 58.6 tok/s - 53.6 tok/s

响应

我是 Claude，由 Anthropic 开发，属于 Claude 模型系列（当前为 Opus 4.8）。

我是 Claude，由 Anthropic 打造的 AI 助手。当前模型是 Claude Opus 4.8，可用于编程、写作、分析等多种任务。

我是 Claude Code，Anthropic 开发的官方命令行 AI 编程助手，基于 Claude Opus 4.8 模型，帮助你完成软件工程和编码任务。

我是 Claude，一个由 Anthropic 开发的人工智能助手，旨在帮助用户完成各类任务。

我是 Kiro，一款由 Anthropic 技术驱动的 AI 编程助手，专注于帮助开发者完成软件工程任务。

步骤 5: knowledge_recall text

时间 10:32:47 ~ 10:32:57 (8.3s) 10:36:45 ~ 10:36:55 (8.9s) 10:35:22 ~ 10:35:32 (9.6s) 10:35:02 ~ 10:35:07 (4.5s) 10:40:16 ~ 10:40:46 (26.6s)

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) 1,813 ms 5,469 ms 1,612 ms 1,157 ms 1,617 ms

首文本(SSE delta) 5,099 ms 7,094 ms - 3,701 ms 1,618 ms

总耗时 7,274 ms 8,214 ms 8,417 ms 3,771 ms 25,615 ms

输入 2 0 2 2 9

输出 695 342 718 320 1,986

缓存创建

4,778

5m: 0 / 1h: 4,778

4,853

5m: 0 / 1h: 4,853

4,279

5m: 0 / 1h: 4,279

4,918

5m: 4,918 / 1h: 0

缓存读取 0 0 0 0 0

输出速度 burst burst 105.5 tok/s burst 82.8 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 10:35:45 ~ 10:35:50 (2.6s) 10:40:00 ~ 10:40:15 (13.8s) 10:39:26 ~ 10:39:46 (18.6s) 10:38:43 ~ 10:38:48 (1.1s) 10:42:53 ~ 10:43:13 (15.9s)

提示词

查看本步 Prompt

Count digits exactly in s. Return {"n":len(s),"c":[count0,count7,count9]}. s=891307590694737464662284183613650851198851657334725650687024235866383815

HTTP 头(TTFB) 1,090 ms 4,821 ms 4,596 ms - 1,661 ms

首文本(SSE delta) 1,091 ms 10,764 ms - - 1,662 ms

总耗时 1,523 ms 10,816 ms 12,683 ms - 15,099 ms

输入 211 0 2 - 28

输出 19 20 775 - 1,722

缓存创建

5m: 0 / 1h: 0

213

5m: 0 / 1h: 213

180

5m: 180 / 1h: 0

缓存读取 0 0 18,543 - 0

输出速度 - - 95.8 tok/s - 128.2 tok/s

响应

{"n":72,"c":[3,6,7]}

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

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

失败

评测失败

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

总计

总输入(含缓存)

132,472

147,916

129,792

仅前 6 步

101,239

总输出

797

450

1,598

402

仅前 6 步

4,002

总缓存创建

36,757

5m: 0 / 1h: 36,757

36,247

5m: 0 / 1h: 36,247

35,760

5m: 0 / 1h: 35,760

仅前 6 步

101,049

5m: 101,049 / 1h: 0

总缓存读取

94,042

108,339

92,596

仅前 6 步

输出速度

49.2 tok/s

76.4 tok/s

97.3 tok/s

burst

仅前 6 步

90.9 tok/s

总执行时间

29.0s

39.5s

48.0s

18.6s

仅前 6 步

1m 5.4s

总墙钟时间

10m 37.9s

含步骤间等待

15m 6.9s

含步骤间等待

14m 36.0s

含步骤间等待

13m 43.9s

含步骤间等待

仅前 6 步

18m 7.0s

含步骤间等待

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.3050

Anthropic: Claude Opus 4.8

>=$0.0112

Anthropic: Claude Opus 4.8 (成本不完整)

$0.3373

Anthropic: Claude Opus 4.8

$0.2870

Anthropic: Claude Opus 4.8 (部分)

$0.4957

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

ID 格式 msg_01X23GEkio99iyoGXBUpefjQ msg_c927bb722b724362a866718967dd9d28 msg_01WFUPDMUcawhEpfQpWRvfut msg_01JnrqYsobE4qWBWacjP5aKV msg_01iUpZ7I42O0lFlhXq1r4Yzp

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_011Cc8424pZFExL3gLGTQ5hG
R-Kiro	server: Photon-Edge x-amzn-requestid: c3f1ae94-4cf6-4228-96d1-7479e5857cd9	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	1781664003739-gz520dq1w2rk8 c3f1ae94-4cf6-4228-96d1-7479e5857cd9
O-Max	server: nginx/1.29.8 x-oneapi-request-id: 202606170239326991241308268d9d6wcmqATEd	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-idbody:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	—
O-Max-L2	cf-ray: a0cea8a219127ddc-LAX server: Photon-Edge set-cookie: _cfuvid=oJCfo1GvtLJ672KlRIufLsKusHJK5pHuULaWMF3rroY-1781663703.377436-1.0.1.1-NhZTE0dpz8Slez24D7ICNiQcMiMaRjfd.fmARBid5A0; 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	1781663703335-a2pc9mtkzshwv req_011Cc83xrpQTmioVLci3yPB8 a0cea8a219127ddc-LAX
R-ultra	cf-ray: a0ceb4229fd57343-NRT server: cloudflare x-new-api-version: v0.0.0 x-oneapi-request-id: 202606170242545917690688268d9d6VBjlVGyt	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	a0ceb4229fd57343-NRT

协议指纹评分（机器自动）

总分 / 100 98 48 88 55

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 8.6 /10 ×20 0.0 /10 ×20 8.0 /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 5.0 /10 ×14 10.0 /10 ×14 — 10.0 /10 ×14

sliding 5m 10.0 /10 ×13 5.0 /10 ×13 10.0 /10 ×13 — 0.0 /10 ×13

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

系统提示纯净 10.0 /10 ×8 未计入 5.0 /10 ×8 — 0.0 /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 0.0 /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 — 0.0 /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 10.0 /10 ×5 — 9.7 /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

流式投递未计入未计入未计入 — 未计入

综合结论

与基线相似度基准（参考） 48 v3.20.0 88 v3.20.0 执行失败，不评分 55 v3.20.0

雷达图例

共 20 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 自由身份 identity_free_clean
6 停止原因 stop_reason_present
7 系统提示纯净 system_prompt_clean
8 Service Tier service_tier_present
9 缓存命中比 cache_hit_ratio_match
10 缓存 TTL 一致 cache_ttl_consistency
11 推理地区 inference_geo_present
12 缓存连续 cache_continuity_intra
13 知识召回 knowledge_recall_match
14 延迟基线 latency_baseline_match
15 原生 msg-ID anthropic_msg_id_format
16 Sliding 缓存 cache_sliding_correctness
17 结构化身份 identity_structured_match
18 中文遵循 instruction_following_lang
19 档位思考量 tier_thinking_volume_match
20 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 3815 ms 5455 ms 5969 ms 2691 ms 10002 ms

输出速度 49.2 tok/s 76.4 tok/s 97.3 tok/s burst 90.9 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 ⓘ

$0.3050

>=$0.0112 (成本不完整)

$0.3373

$0.2870 (部分)

$0.4957

缓存读取占比 ⓘ 71% - 73% 71% (部分) 0%

揭晓

通道

官方基准

服务商 SSSAiCode ↗

通道R-Kiro

服务商 ByteCatCode ↗

通道O-Max

服务商 SSSAiCode ↗

通道O-Max-L2

服务商 FoxCode ↗

通道R-ultra

返回列表