模型对比 | RPDiag

测试说明点击展开 5 步 quick-probe-v1 详解

同一会话连续 5 步，覆盖 5 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。步与步之间随机延迟 1–4 分钟。第 3 步（cutoff）累计跨度 > 5 分钟保住 sliding cache 检测；全流程总跨度 > 6 分钟容纳新增的第 5 步。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 · 5 步骤

2026/06/16 运行日期

claude-haiku-4-5-20251001 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

HongMaCC

u-cc

LingxiCode

M-

TopRouterCN

O-Max

Novaris

O-Max

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 22:03:25 ~ 22:03:30 (3.2s) 22:03:23 ~ 22:03:28 (3.8s) 22:03:21 ~ 22:03:26 (4.1s) 22:03:19 ~ 22:03:24 (4.5s) 22:03:19 ~ 22:03:29 (5.3s)

提示词

查看本步 Prompt

ping. Reply with the single word: pong

HTTP 头(TTFB) 1,809 ms 2,051 ms 2,582 ms 2,930 ms 2,882 ms

首文本(SSE delta) 1,810 ms 2,052 ms 2,584 ms 2,932 ms 2,884 ms

总耗时 2,152 ms 2,052 ms 2,998 ms 3,017 ms 3,682 ms

输入 10 0 10 1,645 10

输出 54 22 59 37 55

缓存创建

5m: 0 / 1h: 0

33,122

5m: 0 / 1h: 33,122

31,487

5m: 0 / 1h: 31,487

33,138

5m: 0 / 1h: 33,138

缓存读取 33,122 0 0 0 0

输出速度 157.9 tok/s - 142.5 tok/s - 68.9 tok/s

响应

pong

pong

pong

pong

pong

步骤 2: identity text

时间 22:06:26 ~ 22:06:36 (6.8s) 22:04:46 ~ 22:04:51 (4.2s) 22:04:29 ~ 22:04:39 (6.1s) 22:05:26 ~ 22:05:36 (6.3s) 22:05:35 ~ 22:05:50 (10.1s)

提示词

查看本步 Prompt

输出 3 行 metadata 描述你自己（当前 API 模型）。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。禁止任何额外字符：
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) 3,004 ms 2,904 ms 3,254 ms 5,484 ms 7,531 ms

首文本(SSE delta) 3,007 ms 2,905 ms 3,255 ms 5,485 ms 7,532 ms

总耗时 5,853 ms 2,906 ms 5,256 ms 5,570 ms 9,239 ms

输入 10 0 10 10 10

输出 355 207 259 353 202

缓存创建

245

5m: 0 / 1h: 245

5m: 0 / 1h: 0

250

5m: 0 / 1h: 250

1,863

5m: 0 / 1h: 1,863

33,397

5m: 0 / 1h: 33,397

缓存读取 33,122 0 33,122 31,487 0

输出速度 124.7 tok/s burst 129.4 tok/s burst 118.3 tok/s

响应

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

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

步骤 3: cutoff text

时间 22:09:51 ~ 22:09:56 (4.0s) 22:06:29 ~ 22:06:35 (3.1s) 22:06:59 ~ 22:07:04 (4.7s) 22:07:17 ~ 22:07:22 (4.0s) 22:07:29 ~ 22:07:34 (4.6s)

提示词

查看本步 Prompt

Knowledge cutoff date? Reply with YYYY-MM only.

HTTP 头(TTFB) 2,278 ms 1,898 ms 3,208 ms 3,108 ms 2,850 ms

首文本(SSE delta) 2,280 ms 1,899 ms 3,209 ms 3,109 ms 2,851 ms

总耗时 2,853 ms 1,899 ms 3,721 ms 3,240 ms 3,577 ms

输入 10 0 10 10 10

输出 64 35 63 75 97

缓存创建

378

5m: 0 / 1h: 378

5m: 0 / 1h: 0

282

5m: 0 / 1h: 282

376

5m: 0 / 1h: 376

5m: 0 / 1h: 61

缓存读取 33,367 0 33,372 33,350 33,397

输出速度 111.7 tok/s burst 123.0 tok/s - 133.6 tok/s

响应

2025-02

2025-02

2025-02

2025-02

2025-02

步骤 4: identity_free text

时间 22:13:28 ~ 22:13:38 (7.5s) 22:08:34 ~ 22:08:39 (3.6s) 22:09:46 ~ 22:09:56 (6.5s) 22:09:53 ~ 22:10:03 (5.8s) 22:09:21 ~ 22:09:41 (16.1s)

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) 1,985 ms 2,506 ms 3,429 ms 4,909 ms 2,435 ms

首文本(SSE delta) 1,986 ms 2,507 ms 3,430 ms 4,910 ms 2,436 ms

总耗时 6,485 ms 2,507 ms 5,689 ms 5,003 ms 15,281 ms

输入 10 0 10 10 10

输出 649 41 382 368 1,962

缓存创建

106

5m: 0 / 1h: 106

5m: 0 / 1h: 0

103

5m: 0 / 1h: 103

115

5m: 0 / 1h: 115

5m: 0 / 1h: 43

缓存读取 33,745 0 33,654 33,726 33,458

输出速度 144.3 tok/s burst 169.1 tok/s burst 152.7 tok/s

响应

我是Claude，由Anthropic公司开发的AI助手。专长是编程、代码分析和软件工程任务。

我是Claude，属于Claude产品，由Anthropic开发。

我是Claude Haiku 4.5，属于Claude AI助手产品家族，由Anthropic公司开发。Claude是一个多功能的AI助手模型系列。

我是Claude Haiku 4.5，由Anthropic公司开发的AI助手。属于Claude产品家族，是一个快速、轻量级的语言模型。

我是Claude Haiku，由Anthropic开发的AI助手，擅长软件工程任务、代码分析和问题解答。

步骤 5: knowledge_recall text

时间 22:15:51 ~ 22:16:06 (14.7s) 22:11:39 ~ 22:11:54 (11.0s) 22:13:45 ~ 22:14:10 (20.6s) 22:13:24 ~ 22:13:39 (13.1s) 22:11:55 ~ 22:12:10 (14.4s)

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) 1,704 ms 9,359 ms 2,829 ms 7,704 ms 2,742 ms

首文本(SSE delta) 1,707 ms 9,360 ms 2,830 ms 7,706 ms 2,743 ms

总耗时 13,859 ms 9,615 ms 19,592 ms 12,323 ms 13,674 ms

输入 10 0 10 10 10

输出 1,825 1,636 2,589 1,488 1,757

缓存创建

3,397

5m: 0 / 1h: 3,397

5m: 0 / 1h: 0

8,757

5m: 0 / 1h: 8,757

8,827

5m: 0 / 1h: 8,827

8,068

5m: 0 / 1h: 8,068

缓存读取 5,721 0 0 0 0

输出速度 150.2 tok/s burst 154.5 tok/s burst 160.7 tok/s

响应

```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

```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

总计

总输入(含缓存) 143,253 0 142,712 142,916 141,612

总输出 2,947 1,941 3,352 2,321 4,073

总缓存创建

4,126

5m: 0 / 1h: 4,126

5m: 0 / 1h: 0

42,514

5m: 0 / 1h: 42,514

42,668

5m: 0 / 1h: 42,668

74,707

5m: 0 / 1h: 74,707

总缓存读取 139,077 0 100,148 98,563 66,855

输出速度 144.4 tok/s burst 152.7 tok/s burst 150.8 tok/s

总执行时间 36.2s 25.7s 42.0s 33.7s 50.6s

总墙钟时间

12m 41.0s

含步骤间等待

8m 31.1s

含步骤间等待

10m 48.8s

含步骤间等待

10m 19.6s

含步骤间等待

8m 51.1s

含步骤间等待

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.0339

Anthropic: Claude Haiku 4.5

>=$0.0097

Anthropic: Claude Haiku 4.5 (成本不完整)

$0.0800

Anthropic: Claude Haiku 4.5

$0.0765

Anthropic: Claude Haiku 4.5

$0.1205

Anthropic: Claude Haiku 4.5

通道指纹

来源步骤 knowledge_recall knowledge_recall knowledge_recall knowledge_recall knowledge_recall

平台 direct (100%) one-api (100%) one-api (100%) direct (100%) one-api (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_01T7iEg2KzndFKCUAjtTvyTG msg_5c6a3376ba5b8a71ae71537a msg_01SxNRA7oTY2Uoj3MYCwZ8Yy msg_01KkHe2Rvxe7vfAnSWxCaC3R msg_01G5yWDFSXn1Te9oPcDAJ6XX

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: 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_011Cc75ba2kQdGJLnRiUoZ17
u-cc	server: nginx x-new-api-version: 1.0.10 x-oneapi-request-id: 202606161411408704629478268d9d6CeiwdnJc	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-new-api-versionhdr: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: a0ca6ac26b46fd4f-NRT server: cloudflare set-cookie: server_name_session=e44607f7628a2aebd93b10e6388b22e1; Max-Age=86400; httponly; path=/ x-oneapi-request-id: 20260616141346943701254phzMejqV	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	a0ca6ac26b46fd4f-NRT
O-Max	server: nginx/1.27.5	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	tr-ew7y4grd-e7l0kkwoo7-077493
O-Max	server: nginx x-oneapi-request-id: 202606161411561885051638268d9d6vQiOLPWL	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	—

协议指纹评分（机器自动）

总分 / 100 100 58 95 93 84

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 0.0 /10 ×20 7.4 /10 ×20 6.9 /10 ×20 0.0 /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 10.0 /10 ×13 5.0 /10 ×13 10.0 /10 ×13 10.0 /10 ×13 7.7 /10 ×13

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

系统提示纯净 10.0 /10 ×8 未计入 10.0 /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 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 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 9.9 /10 ×5 7.5 /10 ×5 8.1 /10 ×5

流式投递 10.0 /10 ×5 未计入 10.0 /10 ×5 未计入 10.0 /10 ×5

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

与基线相似度基准（参考） 58 v3.19.0 95 v3.19.0 93 v3.19.0 84 v3.19.0

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 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 自报身份一致 self_identity_consistency
21 中文遵循 instruction_following_lang
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 6240 ms 3796 ms 7451 ms 5831 ms 9091 ms

输出速度 144.4 tok/s burst 152.7 tok/s burst 150.8 tok/s

完成步骤 5/5 成功 5/5 成功 5/5 成功 5/5 成功 5/5 成功

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.0339

>=$0.0097 (成本不完整)

$0.0800

$0.0765

$0.1205

缓存读取占比 ⓘ 97% - 70% 69% 47%

揭晓

通道

官方基准

服务商 HongMaCC ↗

通道u-cc

服务商 LingxiCode ↗

通道M-

服务商 TopRouterCN ↗

通道O-Max

服务商 Novaris ↗

通道O-Max

返回列表