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

2026/06/19 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

官方基准

FoxCode

R-ultra

Novaris

O-Max

OmniaKey

O-Max

SAIAi

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 23:19:22 ~ 23:19:27 (4.3s) 23:19:16 ~ 23:19:21 (3.3s) 23:19:20 ~ 23:19:25 (2.8s) 23:19:18 ~ 23:19:28 (6.9s) 23:19:14 ~ 23:19:19 (3.5s)

提示词

查看本步 Prompt

ping. Reply with the single word: pong

HTTP 头(TTFB) 2,823 ms 2,081 ms 1,751 ms 3,472 ms 2,370 ms

首文本(SSE delta) 2,825 ms 2,215 ms 1,757 ms 3,473 ms 2,371 ms

总耗时 2,981 ms 2,310 ms 1,952 ms 3,656 ms 2,526 ms

输入 1,454 9 1,492 1,454 1,454

输出 4 2 4 4 4

缓存创建

5m: 0 / 1h: 0

23,551

5m: 23,551 / 1h: 0

27,567

5m: 27,567 / 1h: 0

30,222

5m: 0 / 1h: 30,222

30,222

5m: 0 / 1h: 30,222

缓存读取 30,222 0 0 0 0

响应

pong

pong

pong

pong

pong

步骤 2: identity text

时间 23:21:36 ~ 23:21:41 (3.8s) 23:21:55 ~ 23:22:00 (4.7s) 23:23:06 ~ 23:23:11 (2.9s) 23:21:21 ~ 23:21:31 (9.1s) 23:23:17 ~ 23:23:22 (4.1s)

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。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) 2,408 ms 1,976 ms 1,539 ms 7,271 ms 2,764 ms

首文本(SSE delta) 2,410 ms 1,977 ms 1,551 ms 7,272 ms 2,769 ms

总耗时 2,734 ms 3,778 ms 2,004 ms 7,887 ms 3,083 ms

输入 2 55 2 2 2

输出 28 126 28 28 28

缓存创建

1,646

5m: 0 / 1h: 1,646

160

5m: 160 / 1h: 0

1,684

5m: 1,684 / 1h: 0

1,646

5m: 0 / 1h: 1,646

5m: 0 / 1h: 0

缓存读取 30,222 23,551 27,567 30,222 31,868

输出速度 - 70.0 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-4-8

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

步骤 3: cutoff text

时间 23:24:05 ~ 23:24:10 (3.7s) 23:25:42 ~ 23:25:47 (4.5s) 23:24:56 ~ 23:25:01 (3.1s) 23:22:51 ~ 23:23:01 (6.2s) 23:25:58 ~ 23:26:03 (3.5s)

提示词

查看本步 Prompt

Knowledge cutoff date? Reply with YYYY-MM only.

HTTP 头(TTFB) 2,287 ms 1,969 ms 2,029 ms 4,199 ms 2,243 ms

首文本(SSE delta) 2,289 ms 1,970 ms 2,044 ms 4,200 ms 2,245 ms

总耗时 2,677 ms 3,625 ms 2,332 ms 4,637 ms 2,494 ms

输入 2 99 2 2 2

输出 6 55 6 6 6

缓存创建

5m: 0 / 1h: 54

24,142

5m: 24,142 / 1h: 0

5m: 54 / 1h: 0

5m: 0 / 1h: 54

5m: 0 / 1h: 0

缓存读取 31,868 0 29,251 31,868 31,922

输出速度 - 33.2 tok/s - - -

响应

2026-01

2026-01

2025-03

2026-01

2026-01

步骤 4: identity_free text

时间 23:28:02 ~ 23:28:07 (4.5s) 23:27:28 ~ 23:27:38 (5.1s) 23:28:07 ~ 23:28:12 (4.1s) 23:26:42 ~ 23:26:52 (6.4s) 23:29:16 ~ 23:29:21 (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) 2,334 ms 2,101 ms 2,206 ms 4,173 ms 2,124 ms

首文本(SSE delta) 2,336 ms 2,102 ms 2,216 ms 4,174 ms 2,126 ms

总耗时 3,503 ms 4,335 ms 3,359 ms 5,305 ms 2,694 ms

输入 2 52 2 2 2

输出 67 81 66 61 44

缓存创建

5m: 0 / 1h: 39

431

5m: 431 / 1h: 0

5m: 35 / 1h: 0

5m: 0 / 1h: 31

缓存读取 31,922 24,142 29,305 31,922 31,922

输出速度 57.4 tok/s 36.3 tok/s 57.7 tok/s 53.9 tok/s 77.5 tok/s

响应

我是 Claude，由 Anthropic 开发的 AI 助手。我擅长编程、写作、分析和回答各类问题，致力于提供有帮助、诚实且安全的协助。

我是 Kiro，由 Anthropic 开发的 AI 助手，基于 Claude 模型（claude-opus-4-8）构建，专注于协助软件工程及其他专业任务。

我是 Claude，由 Anthropic 打造的 AI 助手。当前以 Claude Code 命令行工具的形式运行，可协助编程、回答问题及处理各类任务。

我是 Claude，由 Anthropic 开发的 AI 助手。当前模型是 Claude Opus 4.8，可协助编程、研究、写作等多种任务。

我是 Claude，由 Anthropic 开发的人工智能助手，具体模型为 Claude Opus 4.8。

步骤 5: knowledge_recall text

时间 23:30:01 ~ 23:30:11 (9.6s) 23:29:28 ~ 23:29:58 (25.9s) 23:30:45 ~ 23:30:55 (7.0s) 23:28:39 ~ 23:28:54 (11.2s) 23:30:39 ~ 23:30:49 (8.6s)

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) 2,564 ms 2,253 ms 1,859 ms 2,840 ms 2,354 ms

首文本(SSE delta) 6,313 ms 2,254 ms 3,849 ms 5,673 ms 5,174 ms

总耗时 8,687 ms 25,216 ms 6,114 ms 8,151 ms 7,743 ms

输入 2 52 2 2 2

输出 726 2,154 572 686 687

缓存创建

4,782

5m: 0 / 1h: 4,782

4,801

5m: 4,801 / 1h: 0

3,070

5m: 3,070 / 1h: 0

4,768

5m: 0 / 1h: 4,768

2,979

5m: 0 / 1h: 2,979

缓存读取 0 0 0 0 1,772

输出速度 burst 93.8 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 23:32:59 ~ 23:33:04 (2.4s) 23:32:19 ~ 23:32:55 (30.9s) 23:32:23 ~ 23:32:38 (11.1s) 23:32:07 ~ 23:32:17 (5.3s) 23:32:16 ~ 23:32:21 (2.7s)

提示词

查看本步 Prompt

Count digits exactly in s. Return {"n":len(s),"c":[count0,count7,count9]}. s=240148202106814303920831970378351885152488241024404115140340257235923391

HTTP 头(TTFB) 1,085 ms 2,510 ms 1,841 ms 3,211 ms 1,256 ms

首文本(SSE delta) 1,086 ms 2,511 ms 10,128 ms 3,212 ms 1,258 ms

总耗时 1,463 ms 29,998 ms 10,243 ms 4,291 ms 1,730 ms

输入 211 56 193 211 211

输出 19 1,963 1,024 19 19

缓存创建

5m: 0 / 1h: 0

152

5m: 152 / 1h: 0

5m: 0 / 1h: 0

缓存读取 0 0 0 0 0

输出速度 - 71.4 tok/s burst - -

响应

{"n":71,"c":[12,5,8]}

{"n":72,"c":[10,3,4]}

{"n":72,"c":[10,3,4]}

{"n":71,"c":[8,3,8]}

{"n":71,"c":[8,3,5]}

总计

总输入(含缓存) 132,428 101,253 120,226 132,406 132,389

总输出 850 4,381 1,700 804 788

总缓存创建

6,521

5m: 0 / 1h: 6,521

53,237

5m: 53,237 / 1h: 0

32,410

5m: 32,410 / 1h: 0

36,721

5m: 0 / 1h: 36,721

33,232

5m: 0 / 1h: 33,232

总缓存读取 124,234 47,693 86,123 94,012 97,484

输出速度 57.4 tok/s 78.0 tok/s 57.7 tok/s 53.9 tok/s 77.5 tok/s

总执行时间 28.4s 1m 14.4s 30.9s 45.1s 26.1s

总墙钟时间

13m 42.9s

含步骤间等待

13m 39.1s

含步骤间等待

13m 18.9s

含步骤间等待

12m 59.9s

含步骤间等待

13m 7.9s

含步骤间等待

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.1325

Anthropic: Claude Opus 4.8

$0.3429

Anthropic: Claude Opus 4.8

$0.2206

Anthropic: Claude Opus 4.8

$0.3050

Anthropic: Claude Opus 4.8

$0.2845

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_01WovZxjFqrSFLonAA3qYT1g msg_01dflhkONNvIqn7hfCoX2tZd msg_01GMkw8UDioE8ThsdHb1sxbx msg_01JUk7zSkJr4qN8GPd6uWiyJ msg_01VxZKvzgrjMaGsUk9e2P9g9

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_011CcCruVrVFLGh1xNYzw49V
R-ultra	cf-ray: a0e395febfd6e3bf-NRT server: cloudflare x-new-api-version: v0.0.0 x-oneapi-request-id: 20260619153221101263568268d9d6QsKbko2J	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	a0e395febfd6e3bf-NRT
O-Max	server: nginx x-oneapi-request-id: 202606191532247158656228268d9d600MRAQwA	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	cf-ray: a0e395b44b080e30-AMS server: cloudflare x-oneapi-request-id: 202606191532095249165528268d9d6XCrQF9Vy	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	a0e395b44b080e30-AMS
O-Max	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_011CcCrrLAFApT4J7QP2pCXW

协议指纹评分（机器自动）

总分 / 100 100 68 73 95 100

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 10.0 /10 ×20 5.5 /10 ×20 10.0 /10 ×20 10.0 /10 ×20 10.0 /10 ×20

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

sliding 5m 10.0 /10 ×13 7.7 /10 ×13 10.0 /10 ×13 10.0 /10 ×13 10.0 /10 ×13

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

系统提示纯净 10.0 /10 ×8 0.0 /10 ×8 0.0 /10 ×8 10.0 /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 10.0 /10 ×7

知识截止 10.0 /10 ×7 10.0 /10 ×7 0.0 /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 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

inference_geo 10.0 /10 ×5 10.0 /10 ×5 10.0 /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 7.6 /10 ×5 10.0 /10 ×5

流式投递 10.0 /10 ×5 6.4 /10 ×5 10.0 /10 ×5 3.9 /10 ×5 10.0 /10 ×5

Req-ID 透传 10.0 /10 ×4 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 10.0 /10 ×3

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

与基线相似度基准（参考） 68 v3.24.0 73 v3.24.0 95 v3.24.0 100 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 3674 ms 11544 ms 4334 ms 5655 ms 3378 ms

输出速度 57.4 tok/s 78.0 tok/s 57.7 tok/s 53.9 tok/s 77.5 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 ⓘ

$0.1325

$0.3429

$0.2206

$0.3050

$0.2845

缓存读取占比 ⓘ 94% 47% 72% 71% 74%

揭晓

通道

官方基准

服务商 FoxCode ↗

通道R-ultra

服务商 Novaris ↗

通道O-Max

服务商 OmniaKey ↗

通道O-Max

服务商 SAIAi ↗

通道O-Max

返回列表