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

2026/06/24 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

SudoCode

M-Max

Novaris

O-Max

FoxCode

R-ultra

官方基准

AiTokensFlux

O-Pro

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 01:16:59 ~ 01:17:09 (8.2s) 01:16:57 ~ 01:17:02 (4.0s) 01:16:55 ~ 01:17:00 (4.3s) 01:17:01 ~ 01:17:06 (3.8s) 01:16:53 ~ 01:17:38 (44.0s)

提示词

查看本步 Prompt

ping. Reply with the single word: pong

HTTP 头(TTFB) 6,036 ms 1,985 ms 2,372 ms 2,012 ms 43,081 ms

首文本(SSE delta) 6,075 ms 2,007 ms 2,687 ms 2,013 ms 43,083 ms

总耗时 6,581 ms 2,288 ms 2,800 ms 2,151 ms 43,086 ms

输入 1,454 1,492 67 1,454 1,454

输出 4 4 2 4 4

缓存创建

30,222

5m: 0 / 1h: 30,222

27,567

5m: 27,567 / 1h: 0

23,493

5m: 23,493 / 1h: 0

30,222

5m: 0 / 1h: 30,222

30,380

5m: 0 / 1h: 30,380

缓存读取 0 0 0 0 0

响应

pong

pong

pong

pong

pong

步骤 2: identity text

时间 01:19:01 ~ 01:19:11 (5.8s) 01:20:56 ~ 01:21:01 (3.7s) 01:18:13 ~ 01:18:18 (4.5s) 01:20:23 ~ 01:20:28 (3.9s) 01:20:27 ~ 01:20:32 (4.4s)

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) 3,468 ms 2,286 ms 2,048 ms 2,586 ms 2,733 ms

首文本(SSE delta) 3,478 ms 2,312 ms 2,049 ms 2,588 ms 2,733 ms

总耗时 4,601 ms 2,822 ms 3,143 ms 2,968 ms 3,331 ms

输入 2 2 21 2 2

输出 28 30 64 28 28

缓存创建

1,656

5m: 0 / 1h: 1,656

1,694

5m: 1,694 / 1h: 0

260

5m: 260 / 1h: 0

1,656

5m: 0 / 1h: 1,656

1,656

5m: 0 / 1h: 1,656

缓存读取 30,222 27,567 23,493 30,222 30,380

输出速度 - 58.8 tok/s 58.5 tok/s - -

响应

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

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

步骤 3: cutoff text

时间 01:23:01 ~ 01:23:06 (4.4s) 01:24:28 ~ 01:24:33 (4.1s) 01:21:10 ~ 01:21:15 (3.9s) 01:23:54 ~ 01:23:59 (3.6s) 01:21:55 ~ 01:22:00 (4.3s)

提示词

查看本步 Prompt

Knowledge cutoff date? Reply with YYYY-MM only.

HTTP 头(TTFB) 3,004 ms 2,735 ms 2,537 ms 2,367 ms 3,163 ms

首文本(SSE delta) 3,005 ms 2,754 ms 2,538 ms 2,376 ms 3,164 ms

总耗时 3,283 ms 3,047 ms 3,038 ms 2,582 ms 3,404 ms

输入 2 2 68 2 2

输出 6 6 11 6 6

缓存创建

5m: 0 / 1h: 54

5m: 64 / 1h: 0

356

5m: 356 / 1h: 0

5m: 0 / 1h: 64

5m: 0 / 1h: 62

缓存读取 31,878 29,261 23,753 31,878 32,036

响应

2026-01

2025-03

2026-01

2026-01

2026-01

步骤 4: identity_free text

时间 01:25:59 ~ 01:26:09 (6.0s) 01:26:40 ~ 01:26:50 (9.3s) 01:22:32 ~ 01:22:42 (5.5s) 01:27:54 ~ 01:27:59 (4.4s) 01:24:40 ~ 01:24:45 (4.7s)

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) 3,790 ms 7,472 ms 1,628 ms 2,079 ms 2,520 ms

首文本(SSE delta) 3,791 ms 7,493 ms 1,629 ms 2,080 ms 2,529 ms

总耗时 4,599 ms 8,463 ms 4,660 ms 3,287 ms 3,493 ms

输入 2 2 68 2 2

输出 66 61 92 58 59

缓存创建

5m: 0 / 1h: 39

5m: 35 / 1h: 0

357

5m: 357 / 1h: 0

5m: 0 / 1h: 39

5m: 0 / 1h: 35

缓存读取 31,932 29,325 24,109 31,942 32,098

输出速度 81.7 tok/s 62.9 tok/s 30.4 tok/s 48.1 tok/s 61.2 tok/s

响应

你好！我是 Claude，由 Anthropic 开发的 AI 助手。我可以帮你编写代码、解答问题、处理各类软件工程任务。很高兴为你服务！

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，帮助处理编程与各类任务。

我是 Kiro，一款由 Anthropic 开发的AI 编程助手，基于 Claude 模型构建，专注于帮助开发者完成软件工程任务，也支持写作、分析、规划等多种专业工作。

我是 Claude，由 Anthropic 开发的 AI 助手，擅长编程、写作与问答，致力于提供有帮助、安全可靠的协助。

我是 Claude，由 Anthropic 打造的 AI 助手，当前模型为 Claude Opus 4.8，可协助编程、写作、分析等任务。

步骤 5: knowledge_recall text

时间 01:27:13 ~ 01:27:59 (44.7s) 01:28:56 ~ 01:29:06 (8.9s) 01:24:21 ~ 01:25:02 (36.6s) 01:31:46 ~ 01:31:57 (9.1s) 01:25:55 ~ 01:26:06 (9.3s)

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) 4,105 ms 2,567 ms 2,649 ms 2,637 ms 2,671 ms

首文本(SSE delta) 43,700 ms 5,479 ms 2,651 ms 6,572 ms 4,902 ms

总耗时 43,782 ms 7,960 ms 35,609 ms 7,986 ms 8,457 ms

输入 2 2 14 2 2

输出 702 646 2,635 676 671

缓存创建

4,791

5m: 0 / 1h: 4,791

3,085

5m: 3,085 / 1h: 0

4,758

5m: 4,758 / 1h: 0

4,793

5m: 0 / 1h: 4,793

4,788

5m: 0 / 1h: 4,788

缓存读取 0 0 0 0 0

输出速度 burst burst 80.0 tok/s 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 01:30:19 ~ 01:30:24 (3.5s) 01:31:18 ~ 01:31:33 (11.3s) 01:27:23 ~ 01:27:53 (25.3s) 01:34:13 ~ 01:34:18 (2.9s) 01:29:58 ~ 01:30:03 (2.7s)

提示词

查看本步 Prompt

Count digits exactly in s. Return {"n":len(s),"c":[count0,count7,count9]}. s=185724267162898049534800061064976183812343484641797706977312468823453893

HTTP 头(TTFB) 2,112 ms 2,975 ms 2,086 ms 1,597 ms 1,530 ms

首文本(SSE delta) 2,115 ms 10,032 ms 2,087 ms 1,600 ms 1,531 ms

总耗时 2,559 ms 10,594 ms 24,260 ms 1,967 ms 1,933 ms

输入 211 677 89 211 211

输出 19 805 2,443 19 19

缓存创建

5m: 0 / 1h: 0

120

5m: 120 / 1h: 0

5m: 0 / 1h: 0

缓存读取 0 0 0 0 0

输出速度 - burst 110.2 tok/s - -

响应

{"n":70,"c":[5,8,7]}

{"n":72,"c":[6,8,6]}

{"n":72,"c":[6,8,6]}

{"n":70,"c":[5,8,7]}

{"n":70,"c":[5,8,6]}

总计

总输入(含缓存) 132,467 120,775 101,026 132,489 133,108

总输出 825 1,552 5,247 791 787

总缓存创建

36,762

5m: 0 / 1h: 36,762

32,445

5m: 32,445 / 1h: 0

29,344

5m: 29,344 / 1h: 0

36,774

5m: 0 / 1h: 36,774

36,921

5m: 0 / 1h: 36,921

总缓存读取 94,032 86,153 71,355 94,042 94,514

输出速度 81.7 tok/s 61.5 tok/s 88.3 tok/s 48.1 tok/s 61.2 tok/s

总执行时间 1m 12.6s 41.3s 1m 20.0s 27.8s 1m 9.4s

总墙钟时间

13m 25.0s

含步骤间等待

14m 36.0s

含步骤间等待

10m 58.2s

含步骤间等待

17m 17.0s

含步骤间等待

13m 10.0s

含步骤间等待

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.3058

Anthropic: Claude Opus 4.8

$0.2195

Anthropic: Claude Opus 4.8

$0.2831

Anthropic: Claude Opus 4.8

$0.3050

Anthropic: Claude Opus 4.8

$0.3061

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

ID 格式 msg_01JPxLdQHPhuixm4G7H4Tu3B msg_01Q9iZPsts4DDtbdpVbqBSKy msg_01IwmqdmiFWuPh7xATFCsMrm msg_01SrhjxWm27PTjbXDaMVzMBS msg_01LgJ2tFwzWYap9RSQnyMxAg

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-Max	server: nginx x-new-api-version: v1.0.0-rc.14 x-oneapi-request-id: 202606240130202044025567TKTTVlN	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	—
O-Max	server: nginx x-oneapi-request-id: 202606231731188039127538268d9d65Hqzmg3J	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	—
R-ultra	cf-ray: a10534086fb9e041-NRT server: cloudflare x-new-api-version: v0.0.0 x-oneapi-request-id: 202606231727244177644168268d9d6BzXQJ9bx	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	a10534086fb9e041-NRT
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_011CcLbPJjxDsbkVc3ieHFCk
O-Pro	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	—

协议指纹评分（机器自动）

总分 / 100 95 69 75 99 95

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 10.0 /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 10.0 /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 6.7 /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 10.0 /10 ×7

身份(自由) 10.0 /10 ×7 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 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 10.0 /10 ×5 10.0 /10 ×5

流式投递 0.0 /10 ×5 10.0 /10 ×5 8.0 /10 ×5 7.3 /10 ×5 1.4 /10 ×5

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

stop_reason 10.0 /10 ×3 10.0 /10 ×3 10.0 /10 ×3 10.0 /10 ×3 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 95 v3.24.0 69 v3.24.0 75 v3.24.0 基准（参考） 95 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 10901 ms 5862 ms 12252 ms 3490 ms 10617 ms

输出速度 81.7 tok/s 61.5 tok/s 88.3 tok/s 48.1 tok/s 61.2 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 ⓘ

$0.3058

$0.2195

$0.2831

$0.3050

$0.3061

缓存读取占比 ⓘ 71% 71% 71% 71% 71%

揭晓

通道

服务商 SudoCode ↗

通道M-Max

服务商 Novaris ↗

通道O-Max

服务商 FoxCode ↗

通道R-ultra

官方基准

服务商 AiTokensFlux ↗

通道O-Pro

返回列表