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-haiku-4-5-20251001 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

AnyAIGC

M-

FoxCode

官方基准

Novaris

O-Max

LingxiCode

M-

基本信息

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

模型 claude-haiku-4-5-20251001 claude-haiku-4-5-20251001 claude-haiku-4-5-20251001 claude-haiku-4-5-20251001 claude-haiku-4-5-20251001

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 21:39:26 ~ 21:39:31 (4.8s) 21:39:32 ~ 21:39:47 (10.5s) 21:39:34 ~ 21:39:39 (4.3s) 21:39:30 ~ 21:39:35 (4.4s) 21:39:28 ~ 21:39:33 (4.8s)

提示词

查看本步 Prompt

ping. Reply with the single word: pong

HTTP 头(TTFB) 3,122 ms 8,945 ms 2,684 ms 2,900 ms 2,962 ms

首文本(SSE delta) 3,124 ms 9,057 ms 2,686 ms 2,902 ms 2,964 ms

总耗时 3,375 ms 9,594 ms 3,016 ms 3,294 ms 3,360 ms

输入 8,600 10 10 10 10

输出 2 50 56 60 39

缓存创建

33,122

5m: 33,122 / 1h: 0

5m: 0 / 1h: 0

33,122

5m: 33,122 / 1h: 0

33,124

5m: 0 / 1h: 33,124

缓存读取 0 0 33,122 0 0

输出速度 - 93.1 tok/s 169.7 tok/s 153.1 tok/s 98.5 tok/s

响应

pong

pong

pong

pong

pong

步骤 2: identity text

时间 21:41:39 ~ 21:41:49 (5.9s) 21:41:50 ~ 21:42:00 (6.1s) 21:41:34 ~ 21:41:39 (4.3s) 21:41:52 ~ 21:42:28 (33.9s) 21:41:15 ~ 21:41:25 (5.5s)

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) 4,162 ms 3,419 ms 1,988 ms 11,015 ms 2,942 ms

首文本(SSE delta) 4,163 ms 3,420 ms 1,990 ms 32,492 ms 2,943 ms

总耗时 4,182 ms 5,262 ms 3,265 ms 32,921 ms 4,573 ms

输入 8,785 10 10 147 10

输出 25 195 183 21 189

缓存创建

178

5m: 178 / 1h: 0

236

5m: 0 / 1h: 236

3,712

5m: 0 / 1h: 0

33,341

5m: 0 / 1h: 33,341

缓存读取 0 33,122 33,122 3,291 0

输出速度 - 105.9 tok/s 143.5 tok/s - 116.0 tok/s

响应

vendor: anthropic  
brand: anthropic  
model: claude-haiku-4-5-20251001

vendor: anthropic
brand: claude
model: claude-haiku-4-5-20251001

vendor: anthropic
brand: claude
model: claude-haiku-4-5-20251001

vendor: anthropic
brand: claude
model: claude-haiku-4-5-20251001

vendor: anthropic
brand: claude
model: claude-haiku-4-5-20251001

步骤 3: cutoff text

时间 21:45:33 ~ 21:45:48 (10.6s) 21:45:49 ~ 21:45:54 (4.1s) 21:45:05 ~ 21:45:10 (4.0s) 21:45:47 ~ 21:45:57 (5.0s) 21:45:07 ~ 21:45:17 (5.9s)

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) 6,640 ms 2,375 ms 2,397 ms 4,032 ms 4,345 ms

首文本(SSE delta) 6,649 ms 2,377 ms 2,401 ms 4,033 ms 4,346 ms

总耗时 8,961 ms 2,920 ms 2,736 ms 4,297 ms 4,881 ms

输入 8,846 10 10 31 10

输出 5 64 58 2 66

缓存创建

5m: 42 / 1h: 0

206

5m: 0 / 1h: 206

646

5m: 0 / 1h: 0

33,572

5m: 0 / 1h: 33,572

缓存读取 0 33,300 33,358 6,534 0

输出速度 - 117.9 tok/s 173.1 tok/s - 123.4 tok/s

响应

2025-05

2025-02

2025-02

2025-02

2025-02

步骤 4: identity_free text

时间 21:47:07 ~ 21:47:22 (10.4s) 21:48:25 ~ 21:48:30 (4.6s) 21:46:33 ~ 21:46:43 (7.5s) 21:49:39 ~ 21:52:34 (170.8s) 21:49:02 ~ 21:49:12 (6.2s)

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) 8,066 ms 2,437 ms 2,080 ms 2,811 ms 3,621 ms

首文本(SSE delta) 8,067 ms 2,438 ms 2,081 ms 2,812 ms 3,622 ms

总耗时 8,138 ms 3,723 ms 6,529 ms 7,426 ms 4,942 ms

输入 8,887 10 10 10 10

输出 27 192 582 744 192

缓存创建

5m: 35 / 1h: 0

5m: 0 / 1h: 92

33,466

5m: 33,466 / 1h: 0

100

5m: 0 / 1h: 100

缓存读取 0 33,342 33,564 0 33,572

输出速度 - 149.4 tok/s 130.8 tok/s 161.2 tok/s 145.5 tok/s

响应

我是由Anthropic开发的AI助手，提供专业的技术支持和信息解答，帮助用户高效完成任务。

我是Claude，由Anthropic公司开发的AI助手。我可以帮助进行编程、代码分析、软件工程和各种信息处理任务。

我是Claude，由Anthropic公司开发的AI助手。可帮助用户进行代码编写、问题解答和软件工程任务。

我是Claude，由Anthropic开发的AI助手。可以协助编程、文本分析、代码编辑等任务。

我是Claude，由Anthropic公司开发的AI助手。我可以帮助用户进行编程、分析代码、回答问题和执行各种文本任务。

步骤 5: knowledge_recall text

时间 21:50:45 ~ 21:50:55 (9.3s) 21:52:11 ~ 21:52:26 (14.8s) 21:48:44 ~ 21:49:09 (21.7s) 21:53:48 ~ 21:54:08 (18.8s) 21:51:43 ~ 21:52:38 (52.1s)

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) 6,507 ms 1,854 ms 1,598 ms 1,834 ms 2,033 ms

首文本(SSE delta) 6,508 ms 1,855 ms 1,599 ms 1,836 ms 2,034 ms

总耗时 7,530 ms 13,609 ms 20,727 ms 18,093 ms 51,031 ms

输入 9,521 10 10 10 10

输出 154 2,168 2,717 2,377 7,596

缓存创建

8,039

5m: 8,039 / 1h: 0

8,856

5m: 0 / 1h: 8,856

8,828

5m: 0 / 1h: 8,828

8,482

5m: 0 / 1h: 8,482

缓存读取 0 0 0 0 0

输出速度 150.7 tok/s 184.4 tok/s 142.0 tok/s 146.2 tok/s 155.0 tok/s

响应

{"facts":{"super_bowl_lviii":"UNKNOWN","oscars2024_best_picture":"UNKNOWN","ucl_2024":"UNKNOWN","nba_2024":"UNKNOWN","copa_america_2024":"UNKNOWN","euro2024":"UNKNOWN","nobel_peace_2024":"UNKNOWN","wo

```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

```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

```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

```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

步骤 6: digit_count text

时间 21:53:28 ~ 21:53:38 (8.2s) 21:54:44 ~ 21:55:04 (19.8s) 21:52:59 ~ 21:53:39 (38.2s) 21:56:53 ~ 21:57:23 (28.1s) 21:55:56 ~ 21:56:26 (29.0s)

提示词

查看本步 Prompt

Count digits exactly in s. Return {"n":len(s),"c":[count0,count7,count9]}. s=465241139467573957743589140208535693346762620925908637956130978681387505

HTTP 头(TTFB) 5,517 ms 1,569 ms 4,177 ms 1,405 ms 1,640 ms

首文本(SSE delta) 5,825 ms 1,570 ms 4,229 ms 1,406 ms 1,641 ms

总耗时 5,850 ms 19,015 ms 37,238 ms 27,364 ms 28,123 ms

输入 166 193 193 193 193

输出 14 3,051 4,757 3,995 4,821

缓存创建

5m: 0 / 1h: 0

缓存读取 0 0 0 0 0

输出速度 - 174.9 tok/s 144.1 tok/s 153.9 tok/s 182.0 tok/s

响应

{"n":81,"c":[2,1,4]}

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

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

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

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

总计

总输入(含缓存) 44,805 141,423 142,799 90,000 142,434

总输出 227 5,720 8,353 7,199 12,903

总缓存创建

41,416

5m: 41,416 / 1h: 0

9,390

5m: 0 / 1h: 9,390

79,774

5m: 66,588 / 1h: 8,828

108,619

5m: 0 / 1h: 108,619

总缓存读取 0 99,764 133,166 9,825 33,572

输出速度 150.7 tok/s 171.2 tok/s 142.7 tok/s 152.0 tok/s 162.6 tok/s

总执行时间 49.2s 59.8s 1m 20.1s 4m 21.1s 1m 43.6s

总墙钟时间

14m 12.0s

含步骤间等待

15m 32.0s

含步骤间等待

14m 5.1s

含步骤间等待

17m 52.6s

含步骤间等待

16m 58.2s

含步骤间等待

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.0459

Anthropic: Claude Haiku 4.5

$0.0712

Anthropic: Claude Haiku 4.5

$0.0671

Anthropic: Claude Haiku 4.5

$0.1004

Anthropic: Claude Haiku 4.5

$0.2039

Anthropic: Claude Haiku 4.5

通道指纹

来源步骤 digit_count digit_count digit_count digit_count digit_count

平台 direct (50%) direct (100%) direct (100%) one-api (100%) one-api (100%)

上游 - Anthropic 直连 (100%) Anthropic 直连 (100%) Anthropic 直连 (100%) Anthropic 直连 (100%)

CDN - cloudflare - - cloudflare

ID 格式自定义 msg_01Q5HTEEAMBBPmvSnkQAkwY7 msg_01TtUXJrsqRsVNNEkhQWq6Gh msg_015RMJtf1htrADpkD6nQVbhL msg_012Ri48X3T2gGJmbc68uJTAv

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: non_standard has_service_tier: false has_inference_geo: false model_has_date_suffix: true usage_has_cache_fields: true	body:id:non_standardbody:model:date_suffixbody:usage.cache_fields	—
cc	cf-ray: a10c39ea9cf57343-NRT server: cloudflare set-cookie: _cfuvid=8RLgWuxlKLGQnWDp.V0_cXuwlrwrqsXhH357saShj2k-1782309285.9094634-1.0.1.1-IotYFl6EHoUtOSM9.RyL3TuutncIEwXZrilawBaZr8Q; 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: true usage_has_cache_fields: true	hdr:server:cloudflarehdr:cf-raybody:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:model:date_suffixbody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	req_011CcNCTXMB4CAyLnUGyuF67 a10c39ea9cf57343-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: true usage_has_cache_fields: true	body:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:model:date_suffixbody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	req_011CcNCKo2GpWZ1tSp5ZYsDr
O-Max	server: nginx x-oneapi-request-id: 20260624135654267600088268d9d6LmPEvs8E	id_format: msg_prefix extra_fields: ["stop_details"] has_service_tier: true has_inference_geo: true model_has_date_suffix: true usage_has_cache_fields: true	hdr:x-oneapi-request-idbody:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:model:date_suffixbody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	—
M-	cf-ray: a10c3badbdb17963-NRT server: cloudflare set-cookie: server_name_session=56fb82d09f6cfa691ede7bc922d16f50; Max-Age=86400; httponly; path=/ x-oneapi-request-id: 20260624135558350420750XwQ1blFn	id_format: msg_prefix extra_fields: ["stop_details"] has_service_tier: true has_inference_geo: true model_has_date_suffix: true usage_has_cache_fields: true	hdr:x-oneapi-request-idhdr:server:cloudflarehdr:cf-raybody:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:model:date_suffixbody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	a10c3badbdb17963-NRT

协议指纹评分（机器自动）

总分 / 100 35 87 97 71 81

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 7.5 /10 ×20 10.0 /10 ×20 0.0 /10 ×20 0.0 /10 ×20

缓存 TTL 一致 0.0 /10 ×15 0.0 /10 ×15 10.0 /10 ×15 0.0 /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 5.0 /10 ×14 10.0 /10 ×14 10.0 /10 ×14 6.7 /10 ×14 10.0 /10 ×14

sliding 5m 5.0 /10 ×13 10.0 /10 ×13 10.0 /10 ×13 7.7 /10 ×13 5.3 /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 10.0 /10 ×8 10.0 /10 ×8

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

知识截止 5.0 /10 ×7 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 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

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

延迟基线 6.9 /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 0.0 /10 ×5 0.0 /10 ×5 0.0 /10 ×5

Req-ID 透传 0.0 /10 ×4 10.0 /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 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 35 v3.24.1 87 v3.24.1 基准（参考） 71 v3.24.1 81 v3.24.1

雷达图例

共 23 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 中文遵循 instruction_following_lang
22 档位思考量 tier_thinking_volume_match
23 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 6339 ms 9021 ms 12252 ms 15566 ms 16152 ms

输出速度 150.7 tok/s 171.2 tok/s 142.7 tok/s 152.0 tok/s 162.6 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 ⓘ

$0.0459

$0.0712

$0.0671

$0.1004

$0.2039

缓存读取占比 ⓘ 0% 71% 93% 11% 24%

揭晓

通道

服务商 AnyAIGC ↗

通道M-

服务商 FoxCode ↗

通道cc

官方基准

服务商 Novaris ↗

通道O-Max

服务商 LingxiCode ↗

通道M-

返回列表