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-sonnet-4-6 官方基准

3 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-

HongMaCC

u-cc

SAIAi

O-Max

AnyAIGC

M-

官方基准

基本信息

服务名 Claude Code Claude Code Claude Code Claude Code Claude Code

模型 claude-sonnet-4-6 claude-sonnet-4-6 claude-sonnet-4-6 claude-sonnet-4-6 claude-sonnet-4-6

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 19:00:05 ~ 19:00:15 (5.6s) 19:00:09 ~ 19:00:14 (3.5s) 19:00:11 ~ 19:00:16 (4.1s) 19:00:07 ~ 19:00:17 (5.3s) 19:00:13 ~ 19:00:23 (5.0s)

提示词

查看本步 Prompt

ping. Reply with the single word: pong

HTTP 头(TTFB) 3,801 ms 2,162 ms 2,748 ms 3,484 ms 3,673 ms

首文本(SSE delta) 3,844 ms 2,163 ms 2,757 ms 3,563 ms 3,674 ms

总耗时 3,864 ms 2,164 ms 2,928 ms 3,592 ms 3,801 ms

输入 14 0 3 4 3

输出 2 25 5 1 5

缓存创建

32,611

5m: 0 / 1h: 32,611

5m: 0 / 1h: 0

32,318

5m: 0 / 1h: 32,318

31,126

5m: 0 / 1h: 31,126

32,318

5m: 0 / 1h: 32,318

缓存读取 0 0 0 0 0

响应

pong

pong

pong

pong

pong

步骤 2: identity text

时间 19:03:29 ~ 19:03:34 (4.8s) 19:01:24 ~ 19:01:29 (4.1s) 19:01:19 ~ 19:01:24 (4.5s) 19:01:26 ~ 19:01:31 (4.4s) 19:04:22 ~ 19:04:27 (4.2s)

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) 3,329 ms 2,729 ms 2,897 ms 2,791 ms 2,914 ms

首文本(SSE delta) 3,374 ms 2,730 ms 2,946 ms 2,792 ms 2,915 ms

总耗时 3,707 ms 2,730 ms 3,452 ms 3,006 ms 3,273 ms

输入 220 0 3 12 3

输出 17 163 22 14 22

缓存创建

5m: 0 / 1h: 15

5m: 0 / 1h: 0

178

5m: 0 / 1h: 178

31,245

5m: 0 / 1h: 31,245

178

5m: 0 / 1h: 178

缓存读取 32,611 0 32,318 0 32,318

输出速度 - burst - - -

响应

vendor: anthropic
brand: claude
model: claude-sonnet-4-6

vendor: anthropic
brand: claude
model: claude-sonnet-4-6

vendor: anthropic
brand: claude
model: claude-sonnet-4-6

vendor: anthropic
brand: claude
model: claude-sonnet-4-6

vendor: anthropic
brand: claude
model: claude-sonnet-4-6

步骤 3: cutoff text

时间 19:06:00 ~ 19:06:10 (5.7s) 19:02:45 ~ 19:02:56 (6.2s) 19:03:11 ~ 19:03:16 (4.2s) 19:04:54 ~ 19:05:05 (6.7s) 19:05:40 ~ 19:05:46 (3.8s)

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) 4,346 ms 5,028 ms 2,994 ms 4,487 ms 2,590 ms

首文本(SSE delta) 4,347 ms 5,029 ms 2,996 ms 4,629 ms 2,592 ms

总耗时 4,347 ms 5,029 ms 3,134 ms 5,335 ms 2,816 ms

输入 46 0 3 20 3

输出 2 53 8 2 8

缓存创建

32,867

5m: 0 / 1h: 32,867

5m: 0 / 1h: 0

5m: 0 / 1h: 57

31,281

5m: 0 / 1h: 31,281

5m: 0 / 1h: 55

缓存读取 0 0 32,496 0 32,496

输出速度 - burst - - -

响应

2025-08

2025-08

2025-08

2025-08

2025-08

步骤 4: identity_free text

时间 19:10:03 ~ 19:10:13 (6.1s) 19:05:20 ~ 19:05:25 (3.3s) 19:04:40 ~ 19:04:45 (4.0s) 19:07:17 ~ 19:07:27 (5.9s) 19:06:52 ~ 19:06:57 (4.5s)

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) 3,545 ms 2,171 ms 1,903 ms 3,922 ms 2,610 ms

首文本(SSE delta) 3,657 ms 2,172 ms 1,904 ms 3,925 ms 2,612 ms

总耗时 4,887 ms 2,172 ms 2,981 ms 4,566 ms 3,526 ms

输入 46 0 3 28 3

输出 42 50 67 27 66

缓存创建

5m: 0 / 1h: 48

5m: 0 / 1h: 0

5m: 0 / 1h: 43

31,302

5m: 0 / 1h: 31,302

5m: 0 / 1h: 43

缓存读取 32,867 0 32,553 0 32,551

输出速度 34.1 tok/s burst 62.2 tok/s - 72.2 tok/s

响应

我是 Kiro，由 Anthropic 开发的 AI 驱动开发环境，基于 Claude 模型构建，专注于帮助开发者完成软件工程任务。

我是Claude，由Anthropic开发。

我是 Claude，由 Anthropic 开发的 AI 助手。我能帮助你完成写作、分析、编程、问答等各类任务，致力于提供有益、无害且诚实的回答。

我叫 Claude，是由 Anthropic 开发的 AI 助手。当前运行的模型版本为 claude-sonnet-4-6。

我是 Claude，由 Anthropic 开发的 AI 助手。我能帮助完成写作、分析、编程、问答等多种任务，致力于提供安全、有益、诚实的回应。

步骤 5: knowledge_recall text

时间 19:12:30 ~ 19:12:40 (6.9s) 19:07:12 ~ 19:07:27 (10.0s) 19:07:31 ~ 19:07:41 (5.2s) 19:09:26 ~ 19:09:36 (6.8s) 19:08:04 ~ 19:08:14 (6.0s)

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) 2,380 ms 8,472 ms 1,812 ms 3,648 ms 2,553 ms

首文本(SSE delta) 2,768 ms 8,473 ms 1,814 ms 3,649 ms 2,555 ms

总耗时 4,420 ms 8,729 ms 4,200 ms 5,378 ms 5,028 ms

输入 753 0 3 36 3

输出 157 1,614 205 142 205

缓存创建

120

5m: 0 / 1h: 120

5m: 0 / 1h: 0

8,052

5m: 0 / 1h: 8,052

8,304

5m: 0 / 1h: 8,304

8,049

5m: 0 / 1h: 8,049

缓存读取 7,824 0 0 0 0

输出速度 95.0 tok/s burst 85.9 tok/s 82.1 tok/s 82.9 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 19:15:41 ~ 19:15:46 (3.5s) 19:11:10 ~ 19:11:20 (6.0s) 19:10:55 ~ 19:11:15 (15.6s) 19:10:45 ~ 19:11:05 (16.1s) 19:09:21 ~ 19:09:36 (10.9s)

提示词

查看本步 Prompt

Count digits exactly in s. Return {"n":len(s),"c":[count0,count7,count9]}. s=565941839309654862660678187067049132635638253376868273991305236950132387

HTTP 头(TTFB) 1,999 ms 4,808 ms 1,100 ms 1,490 ms 1,107 ms

首文本(SSE delta) 2,066 ms 4,809 ms 1,102 ms 1,491 ms 1,109 ms

总耗时 2,360 ms 4,931 ms 14,442 ms 14,856 ms 9,945 ms

输入 206 0 164 145 164

输出 8 687 1,388 195 921

缓存创建

5m: 0 / 1h: 0

缓存读取 0 0 0 0 0

输出速度 - burst 104.0 tok/s 14.6 tok/s 104.2 tok/s

响应

{"n":71,"c":[3,5,4]}

{"n":69,"c":[6,6,7]}

I'll carefully parse **s = `565941839309654862660678187067049132635638253376868273991305236950132387`**

**Step 1 — Length**

Splitting into groups of 10:
```
5659418393 | 0965486266 | 0678187067 | 04

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

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

**Verification:**

String: `565941839309654862660678187067049132635638253376868273991305236950132387`

- **Length**: 72 characters
- **0s** (×6): positions 11,21,28,3

总计

总输入(含缓存) 140,248 0 138,194 133,503 138,187

总输出 228 2,592 1,695 381 1,227

总缓存创建

65,661

5m: 0 / 1h: 65,661

5m: 0 / 1h: 0

40,648

5m: 0 / 1h: 40,648

133,258

5m: 0 / 1h: 133,258

40,643

5m: 0 / 1h: 40,643

总缓存读取 73,302 0 97,367 0 97,365

输出速度 69.0 tok/s burst 98.8 tok/s 22.3 tok/s 97.5 tok/s

总执行时间 32.6s 33.1s 37.5s 45.3s 34.5s

总墙钟时间

15m 41.0s

含步骤间等待

11m 11.0s

含步骤间等待

11m 4.0s

含步骤间等待

10m 57.9s

含步骤间等待

9m 23.0s

含步骤间等待

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.2755

Anthropic: Claude Sonnet 4.6

>=$0.0389

Anthropic: Claude Sonnet 4.6 (成本不完整)

$0.2076

Anthropic: Claude Sonnet 4.6

$0.5062

Anthropic: Claude Sonnet 4.6

$0.2006

Anthropic: Claude Sonnet 4.6

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%) direct (100%) direct (75%) direct (100%)

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

CDN cloudflare - - - -

ID 格式 msg_iPinitVdWgSE9cWdlEzToLtZ msg_d19dfac87cf0c536e543d6d8 msg_014hMZvAtbh5Ei8LXZKjncvF msg_a8402e7d6e71485687eb1349367595f5 msg_01R2nmTtPgZB96E3L2g5UDrQ

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: a10b50f0ca219726-AMS server: cloudflare	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:server:cloudflarehdr:cf-raybody:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	31fa12c7-6116-47fa-b100-c0a3e99d7430 a10b50f0ca219726-AMS
u-cc	server: nginx x-new-api-version: 1.0.15 x-oneapi-request-id: 202606241111116911385108268d9d6BvRQHLxu	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/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_011CcMyxys9szfTcHVSjRhUH
M-	server: nginx	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: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_011CcMyr3ZTeujKkDgzuW8s4

协议指纹评分（机器自动）

总分 / 100 80 61 100 58 100

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 8.1 /10 ×20 0.0 /10 ×20 10.0 /10 ×20 0.0 /10 ×20 10.0 /10 ×20

缓存 TTL 一致 10.0 /10 ×15 0.0 /10 ×15 10.0 /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 5.0 /10 ×14 10.0 /10 ×14 10.0 /10 ×14 10.0 /10 ×14

sliding 5m 7.7 /10 ×13 5.0 /10 ×13 10.0 /10 ×13 0.0 /10 ×13 10.0 /10 ×13

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

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

身份(自由) 0.0 /10 ×7 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 10.0 /10 ×6 0.0 /10 ×6 10.0 /10 ×6

inference_geo 10.0 /10 ×5 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 10.0 /10 ×5

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

Req-ID 透传 0.0 /10 ×4 0.0 /10 ×4 10.0 /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 80 v3.24.1 61 v3.24.1 100 v3.24.1 58 v3.24.1 基准（参考）

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 流式投递 buffer_dump_match
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 信封自报一致 envelope_self_report_match
19 中文遵循 instruction_following_lang
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 3931 ms 4293 ms 5190 ms 6122 ms 4732 ms

输出速度 69.0 tok/s burst 98.8 tok/s 22.3 tok/s 97.5 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 ⓘ

$0.2755

>=$0.0389 (成本不完整)

$0.2076

$0.5062

$0.2006

缓存读取占比 ⓘ 52% - 70% 0% 70%

揭晓

通道

服务商 WorldBase.ai ↗

通道M-

服务商 HongMaCC ↗

通道u-cc

服务商 SAIAi ↗

通道O-Max

服务商 AnyAIGC ↗

通道M-

官方基准

返回列表