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

2026/06/27 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

AnyAIGC

M-

官方基准

SSSAiCode

O-Max

SSSAiCode

R-Kiro

LinkAPI

O-Max

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 15:53:51 ~ 15:54:01 (9.0s) 15:53:59 ~ 15:54:09 (8.0s) 15:53:57 ~ 15:54:02 (3.4s) 15:53:53 ~ 15:54:03 (6.7s) 15:53:55 ~ 15:54:05 (6.3s)

提示词

查看本步 Prompt

ping. Reply with the single word: pong

HTTP 头(TTFB) 2,278 ms 6,740 ms 2,232 ms 2,680 ms 3,955 ms

首文本(SSE delta) 2,279 ms 6,741 ms 2,476 ms 2,704 ms 3,975 ms

总耗时 2,279 ms 6,838 ms 2,480 ms 2,818 ms 4,168 ms

输入 1,454 1,454 1,428 0 1,454

输出 4 4 4 4 4

缓存创建

30,222

5m: 0 / 1h: 30,222

30,222

5m: 0 / 1h: 30,222

29,764

5m: 0 / 1h: 29,764

30,382

5m: 30,382 / 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 15:57:24 ~ 15:57:30 (4.5s) 15:55:49 ~ 15:56:00 (5.4s) 15:57:19 ~ 15:57:24 (3.5s) 15:56:15 ~ 15:56:20 (3.6s) 15:57:54 ~ 15:57:59 (3.1s)

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) 2,592 ms 4,014 ms 1,926 ms 2,780 ms 1,740 ms

首文本(SSE delta) 2,593 ms 4,016 ms 2,294 ms 2,803 ms 1,742 ms

总耗时 3,372 ms 4,376 ms 2,420 ms 2,918 ms 2,168 ms

输入 2 2 2 0 2

输出 28 28 30 28 28

缓存创建

1,630

5m: 0 / 1h: 1,630

1,630

5m: 0 / 1h: 1,630

1,604

5m: 0 / 1h: 1,604

1,630

5m: 1,630 / 1h: 0

缓存读取 30,222 30,222 29,764 0 30,382

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-4-20250514

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 16:00:53 ~ 16:01:03 (5.2s) 15:57:03 ~ 15:57:13 (5.4s) 15:58:43 ~ 15:58:48 (3.1s) 15:59:18 ~ 15:59:23 (4.1s) 16:00:25 ~ 16:00:30 (3.9s)

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) 3,990 ms 3,812 ms 2,212 ms 2,933 ms 2,608 ms

首文本(SSE delta) 3,991 ms 3,814 ms 2,438 ms 3,184 ms 2,625 ms

总耗时 3,991 ms 3,982 ms 2,444 ms 3,252 ms 2,968 ms

输入 2 2 2 0 2

输出 6 6 6 8 6

缓存创建

5m: 0 / 1h: 62

5m: 0 / 1h: 54

5m: 0 / 1h: 66

5m: 64 / 1h: 0

缓存读取 31,852 31,852 31,368 0 32,012

响应

2026-01

2026-01

2025-01

2026-01

2026-01

步骤 4: identity_free text

时间 16:03:05 ~ 16:03:15 (8.0s) 15:58:31 ~ 15:58:41 (5.7s) 16:02:48 ~ 16:02:58 (5.3s) 16:03:19 ~ 16:03:24 (4.3s) 16:03:58 ~ 16:04:03 (4.1s)

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) 3,635 ms 4,098 ms 2,245 ms 3,016 ms 2,083 ms

首文本(SSE delta) 3,636 ms 4,099 ms 3,604 ms 3,372 ms 2,084 ms

总耗时 5,795 ms 4,731 ms 4,527 ms 3,533 ms 3,160 ms

输入 2 2 2 0 2

输出 63 46 66 40 70

缓存创建

5m: 0 / 1h: 35

5m: 0 / 1h: 31

5m: 0 / 1h: 39

5m: 39 / 1h: 0

缓存读取 31,914 31,906 31,434 0 32,076

输出速度 29.2 tok/s 72.8 tok/s 71.5 tok/s - 65.1 tok/s

响应

我是 Claude，由 Anthropic 打造的 AI 助手。当前模型为 Claude Opus 4.8，可协助完成编程、写作、分析等多种任务。

我是 Claude，一个由 Anthropic 开发的人工智能助手。具体模型是 Claude Opus 4.8。

我是 Claude，由 Anthropic 开发的 AI 助手。我可以协助编程、写作、分析和回答各类问题，致力于提供有用、准确且安全的帮助。

我是 Claude，属于 Claude 系列模型（具体为 Opus 4.8），由 Anthropic 开发。

我是 Claude Code，Anthropic 公司开发的官方命令行 AI 助手，基于 Claude Opus 4.8 模型，帮助用户完成软件工程和编程相关任务。

步骤 5: knowledge_recall text

时间 16:05:08 ~ 16:05:28 (15.0s) 16:01:26 ~ 16:01:41 (10.9s) 16:05:26 ~ 16:05:36 (5.4s) 16:06:10 ~ 16:06:25 (10.7s) 16:05:09 ~ 16:05:24 (11.5s)

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) 8,180 ms 4,131 ms 1,758 ms 7,054 ms 4,487 ms

首文本(SSE delta) 10,149 ms 5,544 ms 4,375 ms 9,526 ms 5,959 ms

总耗时 13,709 ms 9,826 ms 4,608 ms 9,964 ms 10,274 ms

输入 2 2 2 0 2

输出 683 656 317 342 690

缓存创建

4,766

5m: 0 / 1h: 4,766

4,737

5m: 0 / 1h: 4,737

4,293

5m: 0 / 1h: 4,293

6,774

5m: 6,774 / 1h: 0

缓存读取 0 0 0 0 0

输出速度 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 16:08:21 ~ 16:08:26 (3.3s) 16:03:12 ~ 16:03:22 (7.0s) 16:07:22 ~ 16:07:27 (1.0s) 16:10:14 ~ 16:10:29 (10.9s) 16:07:33 ~ 16:07:48 (13.2s)

提示词

查看本步 Prompt

Count digits exactly in s. Return {"n":len(s),"c":[count0,count7,count9]}. s=978637014566903566912745451011406031706857535773446294638296056281954111

HTTP 头(TTFB) 2,072 ms 5,661 ms - 5,156 ms 4,997 ms

首文本(SSE delta) 2,073 ms 5,666 ms - 9,926 ms 6,370 ms

总耗时 2,134 ms 5,927 ms - 10,085 ms 12,136 ms

输入 211 211 - 0 2

输出 19 19 - 20 544

缓存创建

5m: 0 / 1h: 0

2,205

5m: 2,205 / 1h: 0

缓存读取 0 0 - 0 0

输出速度 - - - - 94.3 tok/s

响应

{"n":72,"c":[7,6,5]}

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

失败

评测失败

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

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

总计

总输入(含缓存)

132,376

132,327

129,768

仅前 6 步

137,028

总输出

803

759

423

仅前 6 步

442

1,342

总缓存创建

36,715

5m: 0 / 1h: 36,715

36,674

5m: 0 / 1h: 36,674

35,766

5m: 0 / 1h: 35,766

仅前 6 步

41,094

5m: 41,094 / 1h: 0

总缓存读取

93,988

93,980

92,566

仅前 6 步

94,470

输出速度

29.2 tok/s

72.8 tok/s

71.5 tok/s

仅前 6 步

89.7 tok/s

总执行时间

45.1s

42.2s

21.7s

仅前 6 步

40.3s

42.1s

总墙钟时间

14m 35.2s

含步骤间等待

9m 23.0s

含步骤间等待

13m 30.0s

含步骤间等待

仅前 6 步

16m 36.0s

含步骤间等待

13m 53.1s

含步骤间等待

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.3049

Anthropic: Claude Opus 4.8

$0.3035

Anthropic: Claude Opus 4.8

$0.2876

Anthropic: Claude Opus 4.8 (部分)

>=$0.0111

Anthropic: Claude Opus 4.8 (成本不完整)

$0.2486

Anthropic: Claude Opus 4.8

通道指纹

来源步骤 digit_count digit_count knowledge_recall digit_count digit_count

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

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

CDN - - - - -

ID 格式 msg_01EBmR8cfK3Xst2YcgbSu2Xy msg_01TLcDifkXCPDXLurqx7UCdZ msg_019LLcAoCCKbz3St3cno5cae msg_dc2ce8a43d3a4f879ccb7dcd17a7391e msg_01GJS41ZptnmPkyR2DnaBuRg

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
M-	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	—
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_011CcTR5ZzLqsm46toxUQuLY
O-Max	cf-ray: a122f26069174fb1-LAX server: Photon-Edge set-cookie: _cfuvid=ZGxglaUtrFxiUKgkedIVr1LlNtW3VNfq73D7Qplf1KI-1782547527.7438629-1.0.1.1-mq3HxG8h339SCfGHOyP74yED_trLOuRB3nNdlQasweE; 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	1782547527357-z30r8kvp59r8a req_011CcTRFMghKd3f6D7LhXaFR a122f26069174fb1-LAX
R-Kiro	server: Photon-Edge x-amzn-requestid: afe835a4-7de8-4be5-b0d5-5dcaa74e8b90	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	1782547815489-yn545rtunwps6 afe835a4-7de8-4be5-b0d5-5dcaa74e8b90
O-Max	server: nginx set-cookie: server_name_session=7d88fca8b1426ca76c58d5aeba6fb224; Max-Age=86400; httponly; path=/ x-new-api-version: v1.0.0-rc.15 x-oneapi-request-id: 202606270807348827776988268d9d6l9PTrq9W	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	—

协议指纹评分（机器自动）

总分 / 100 96 100 50 84

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 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 — 0.0 /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 — 5.0 /10 ×14 10.0 /10 ×14

sliding 5m 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 — 0.0 /10 ×8 10.0 /10 ×8

系统提示纯净 10.0 /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

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

流式投递 4.1 /10 ×5 10.0 /10 ×5 — 0.0 /10 ×5 10.0 /10 ×5

Req-ID 透传 0.0 /10 ×4 10.0 /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 96 v3.25.0 基准（参考）执行失败，不评分 50 v3.25.0 84 v3.25.0

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 5213 ms 5947 ms 3296 ms 5428 ms 5812 ms

输出速度 29.2 tok/s 72.8 tok/s 71.5 tok/s - 89.7 tok/s

完成步骤 6/6 成功 6/6 成功 5/6 成功 · 1 失败 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 ⓘ

$0.3049

$0.3035

$0.2876 (部分)

>=$0.0111 (成本不完整)

$0.2486

缓存读取占比 ⓘ 71% 71% 71% (部分) - 69%

揭晓

通道

服务商 AnyAIGC ↗

通道M-

官方基准

服务商 SSSAiCode ↗

通道O-Max

服务商 SSSAiCode ↗

通道R-Kiro

服务商 LinkAPI ↗

通道O-Max

返回列表