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

2026/07/01 运行日期

claude-opus-4-8 官方基准

4 通道分数 < 70 与基线相似度

已揭晓揭晓进度

步骤 / 指标

WorldBase.ai

M-

AiTokensFlux

O-Pro

LinkAPI

O-Max

AIMZ

M-Max

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 06:05:24 ~ 06:05:34 (5.0s) 06:05:20 ~ 06:05:30 (5.7s) 06:05:22 ~ 06:05:27 (3.9s) 06:05:18 ~ 06:05:23 (3.5s) 06:05:26 ~ 06:05:31 (3.6s)

提示词

查看本步 Prompt

ping. Reply with the single word: pong

HTTP 头(TTFB) 3,365 ms 4,580 ms 2,690 ms 2,331 ms 2,168 ms

首文本(SSE delta) 3,366 ms 4,623 ms 2,691 ms 2,427 ms 2,170 ms

总耗时 3,400 ms 4,760 ms 2,933 ms 2,510 ms 2,331 ms

输入 24 2,326 5,880 2,300 2,326

输出 2 4 6 4 4

缓存创建

9,730

14,443

5m: 0 / 1h: 14,443

10,917

5m: 10,917 / 1h: 0

1,223

5m: 0 / 1h: 1,223

1,711

5m: 0 / 1h: 1,711

缓存读取 6,253 0 0 12,732 12,732

响应

pong

pong

pong

pong

pong

步骤 2: identity text

时间 06:09:11 ~ 06:09:21 (8.0s) 06:07:09 ~ 06:08:19 (66.0s) 06:08:18 ~ 06:08:23 (4.7s) 06:06:45 ~ 06:06:50 (4.1s) 06:09:07 ~ 06:09:12 (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。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) 6,847 ms 3,919 ms 3,473 ms 2,134 ms 1,941 ms

首文本(SSE delta) 6,848 ms 3,936 ms 3,474 ms 2,952 ms 1,942 ms

总耗时 6,916 ms 4,383 ms 3,831 ms 3,015 ms 2,277 ms

输入 1 2 5,971 2 2

输出 26 28 68 70 28

缓存创建

9,926

2,518

5m: 0 / 1h: 2,518

4,593

5m: 4,593 / 1h: 0

2,492

5m: 0 / 1h: 2,492

2,518

5m: 0 / 1h: 2,518

缓存读取 6,380 14,443 6,491 13,955 14,443

响应

vendor: anthropic
brand: claude
model: claude-opus-4-1-20250805

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

I'm Claude, an AI assistant made by Anthropic.

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 06:11:08 ~ 06:11:18 (5.0s) 06:12:02 ~ 06:12:07 (4.5s) 06:11:27 ~ 06:11:32 (4.0s) 06:10:17 ~ 06:10:22 (4.0s) 06:11:13 ~ 06:11:18 (4.3s)

提示词

查看本步 Prompt

Knowledge cutoff date? Reply with YYYY-MM only.

HTTP 头(TTFB) 3,821 ms 3,369 ms 2,901 ms 2,661 ms 3,168 ms

首文本(SSE delta) 3,822 ms 3,370 ms 2,902 ms 2,821 ms 3,170 ms

总耗时 3,891 ms 3,593 ms 3,067 ms 2,919 ms 3,342 ms

输入 57 2 5,994 2 2

输出 3 6 10 6 6

缓存创建

9,963

17,031

5m: 0 / 1h: 17,031

3,474

5m: 3,474 / 1h: 0

104

5m: 0 / 1h: 104

5m: 0 / 1h: 54

缓存读取 6,403 0 7,651 16,447 16,961

响应

2025-03

2026-01

2025-01

2025-01

2026-01

步骤 4: identity_free text

时间 06:14:24 ~ 06:14:34 (7.3s) 06:13:36 ~ 06:13:47 (6.0s) 06:14:52 ~ 06:14:57 (4.5s) 06:13:38 ~ 06:13:43 (4.8s) 06:12:51 ~ 06:12:56 (4.5s)

提示词

查看本步 Prompt

请用中文介绍你自己是什么产品或助手，由谁开发，50 字以内。

HTTP 头(TTFB) 5,833 ms 3,974 ms 3,239 ms 2,743 ms 2,082 ms

首文本(SSE delta) 5,834 ms 3,975 ms 3,240 ms 3,360 ms 2,083 ms

总耗时 6,146 ms 5,128 ms 3,470 ms 3,592 ms 3,499 ms

输入 1 2 6,009 2 2

输出 70 74 70 47 69

缓存创建

10,002

5m: 0 / 1h: 39

11,156

5m: 11,156 / 1h: 0

5m: 0 / 1h: 37

5m: 0 / 1h: 39

缓存读取 6,428 17,031 0 16,551 17,015

输出速度 - 64.2 tok/s - - 48.7 tok/s

响应

我是 Claude，由 Anthropic 开发的 AI 助手。在当前环境中，我以 Claude Code 的形式工作，帮助开发者编写代码、解决问题、完成各类技术与专业任务。

我是 Claude，由 Anthropic 开发的 AI 助手。我可以帮你写代码、分析问题、研究信息和完成各类软件工程任务。当前基于 Opus 4.8 模型。

我是 Claude，由 Anthropic 开发的 AI 助手。

我是 Claude，由 Anthropic 开发，属于 Claude 系列 AI 助手产品中的 Opus 模型。

我是 Claude，由 Anthropic 开发的 AI 助手。当前为 Claude Code 命令行代理，可协助软件工程任务，如编写、调试代码和管理项目文件。

步骤 5: knowledge_recall text

时间 06:17:31 ~ 06:17:46 (13.1s) 06:15:05 ~ 06:15:15 (9.4s) 06:16:17 ~ 06:16:27 (5.5s) 06:15:22 ~ 06:15:32 (6.4s) 06:15:52 ~ 06:16:02 (9.4s)

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) 6,027 ms 1,942 ms 4,317 ms 2,304 ms 1,982 ms

首文本(SSE delta) 6,044 ms 4,285 ms 4,318 ms 2,811 ms 3,366 ms

总耗时 12,051 ms 8,656 ms 4,560 ms 5,414 ms 8,396 ms

输入 1 2 2,192 2 2

输出 301 712 683 344 736

缓存创建

2,640

5,775

5m: 0 / 1h: 5,775

2,718

5m: 2,718 / 1h: 0

5,272

5m: 0 / 1h: 5,272

5,760

5m: 0 / 1h: 5,760

缓存读取 1,696 0 1,349 0 0

输出速度 50.1 tok/s burst - 132.2 tok/s 146.3 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

```json
{"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

{"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 06:19:59 ~ 06:20:19 (16.3s) 06:17:51 ~ 06:18:06 (12.0s) 06:20:06 ~ 06:20:11 (3.1s) 06:19:02 ~ 06:19:07 (3.5s) 06:17:54 ~ 06:17:59 (3.8s)

提示词

查看本步 Prompt

Count digits exactly in s. Return {"n":len(s),"c":[count0,count7,count9]}. s=595232639178910463334050535210583549398352904524832278418885780412795962

HTTP 头(TTFB) 6,403 ms 10,815 ms 1,387 ms 1,796 ms 2,460 ms

首文本(SSE delta) 6,404 ms 10,819 ms 1,388 ms 2,167 ms 2,461 ms

总耗时 15,170 ms 11,208 ms 2,097 ms 2,274 ms 2,783 ms

输入 1 211 211 140 211

输出 128 19 19 19 19

缓存创建

154

5m: 0 / 1h: 0

缓存读取 97 0 0 0 0

输出速度 14.6 tok/s - - - -

响应

{"n":72,"c":[6,4,8]}

{"n":70,"c":[3,2,8]}

{"n":70,"c":[4,4,8]}

{"n":70,"c":[5,4,7]}

{"n":70,"c":[7,4,9]}

总计

总输入(含缓存) 69,757 73,825 74,606 71,261 73,778

总输出 530 843 856 490 862

总缓存创建

42,415

39,806

5m: 0 / 1h: 39,806

32,858

5m: 32,858 / 1h: 0

9,128

5m: 0 / 1h: 9,128

10,082

5m: 0 / 1h: 10,082

总缓存读取 27,257 31,474 15,491 59,685 61,151

输出速度 29.0 tok/s 64.2 tok/s - 132.2 tok/s 124.9 tok/s

总执行时间 54.6s 1m 43.6s 25.7s 26.3s 28.8s

总墙钟时间

14m 55.1s

含步骤间等待

12m 46.1s

含步骤间等待

14m 48.9s

含步骤间等待

13m 49.1s

含步骤间等待

12m 32.9s

含步骤间等待

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.1930

Anthropic: Claude Opus 4.8

$0.2983

Anthropic: Claude Opus 4.8

$0.2888

Anthropic: Claude Opus 4.8

$0.1114

Anthropic: Claude Opus 4.8

$0.1279

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

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

CDN cloudflare - - cloudflare -

ID 格式 msg_b131ec202e7c4700bca5695878a2ce19 msg_01JMND1eBTwRRhsRbY7uQcJa msg_01LkHYnf2byXbdE6MHaYvoMK msg_01GmD1KyBwrVp5DwE2BkQb65 msg_01FZ4RdSpsitZoHL6fi27mts

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-	cf-ray: a1408e46e9841c0a-AMS server: cloudflare	id_format: msg_prefix has_service_tier: false has_inference_geo: false model_has_date_suffix: false usage_has_cache_fields: true	hdr:server:cloudflarehdr:cf-raybody:id:msg_prefixbody:usage.cache_fieldsupstream:anthropic-direct-from-body	db24e84e-022c-498f-bb6e-0f2a2c72fc05 a1408e46e9841c0a-AMS
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	—
O-Max	server: nginx set-cookie: server_name_session=48b449e29b720b742b1fbf23591a9a4f; Max-Age=86400; httponly; path=/ x-new-api-version: v1.0.0-rc.15 x-oneapi-request-id: 202606302220072521002088268d9d6cNFCYHEp	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	—
M-Max	cf-ray: a1408ce4186afeb9-AMS server: cloudflare x-new-api-version: v0.0.0 x-oneapi-request-id: 202606302219037462480538268d9d6RWJQsgxj	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-idhdr:server:cloudflarehdr:cf-raybody:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	a1408ce4186afeb9-AMS
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_011CcaDgRBNRhNhspcxHPo7U

协议指纹评分（机器自动）

总分 / 100 46 81 62 83 100

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 0.0 /10 ×20 0.5 /10 ×20 0.0 /10 ×20 10.0 /10 ×20 10.0 /10 ×20

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

延迟基线 7.3 /10 ×15 7.8 /10 ×15 9.9 /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 7.7 /10 ×13 10.0 /10 ×13 10.0 /10 ×13

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

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

身份(结构) 6.7 /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 10.0 /10 ×7 10.0 /10 ×7 10.0 /10 ×7 10.0 /10 ×7

service_tier 0.0 /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 0.0 /10 ×5 0.0 /10 ×5 10.0 /10 ×5

inference_geo 0.0 /10 ×5 10.0 /10 ×5 10.0 /10 ×5 10.0 /10 ×5 10.0 /10 ×5

流式投递 6.8 /10 ×5 6.3 /10 ×5 0.0 /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 46 v3.26.0 81 v3.26.0 62 v3.26.0 83 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 7929 ms 6288 ms 3326 ms 3287 ms 3771 ms

输出速度 29.0 tok/s 64.2 tok/s - 132.2 tok/s 124.9 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 ⓘ

$0.1930

$0.2983

$0.2888

$0.1114

$0.1279

缓存读取占比 ⓘ 39% 43% 21% 84% 83%

揭晓

通道

服务商 WorldBase.ai ↗

通道M-

服务商 AiTokensFlux ↗

通道O-Pro

服务商 LinkAPI ↗

通道O-Max

服务商 AIMZ ↗

通道M-Max

官方基准

返回列表