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

2026/06/25 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

0-0

O-Max

HongMaCC

u-cc

官方基准

FoxCode

R-ultra

PackyCode

cc-Zeus

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 20:40:15 ~ 20:40:25 (6.2s) 20:40:11 ~ 20:40:16 (4.3s) 20:40:19 ~ 20:40:24 (4.1s) 20:40:13 ~ 20:40:23 (7.6s) 20:40:17 ~ 20:40:22 (3.7s)

提示词

查看本步 Prompt

ping. Reply with the single word: pong

HTTP 头(TTFB) 3,108 ms 2,921 ms 2,362 ms 4,972 ms 1,778 ms

首文本(SSE delta) 3,110 ms 2,922 ms 2,364 ms 4,974 ms 1,787 ms

总耗时 3,849 ms 2,922 ms 2,680 ms 6,467 ms 2,360 ms

输入 10 0 10 33,019 10

输出 61 22 52 69 43

缓存创建

33,192

5m: 33,192 / 1h: 0

5m: 0 / 1h: 0

33,122

5m: 0 / 1h: 33,122

5m: 0 / 1h: 0

33,123

5m: 0 / 1h: 33,123

缓存读取 0 0 0 0 0

输出速度 82.5 tok/s - 164.6 tok/s 46.2 tok/s 75.0 tok/s

响应

pong

pong

pong

pong

pong

步骤 2: identity text

时间 20:43:51 ~ 20:43:56 (3.9s) 20:44:08 ~ 20:44:13 (4.6s) 20:42:43 ~ 20:42:49 (4.7s) 20:41:30 ~ 20:41:40 (6.3s) 20:43:42 ~ 20:43:47 (4.2s)

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) 2,546 ms 2,963 ms 2,159 ms 2,240 ms 1,408 ms

首文本(SSE delta) 2,547 ms 2,964 ms 2,160 ms 2,241 ms 1,409 ms

总耗时 2,765 ms 2,965 ms 3,592 ms 4,459 ms 3,169 ms

输入 3 0 10 33,595 10

输出 26 197 174 279 185

缓存创建

7,629

5m: 7,629 / 1h: 0

5m: 0 / 1h: 0

243

5m: 0 / 1h: 243

5m: 0 / 1h: 0

234

5m: 0 / 1h: 234

缓存读取 25,629 0 33,122 0 33,123

输出速度 - burst 121.5 tok/s 125.8 tok/s 105.1 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 20:45:15 ~ 20:45:25 (5.3s) 20:47:45 ~ 20:48:00 (13.4s) 20:44:06 ~ 20:44:11 (3.3s) 20:45:38 ~ 20:45:58 (18.3s) 20:45:54 ~ 20:46:00 (4.4s)

提示词

查看本步 Prompt

Knowledge cutoff date? Reply with YYYY-MM only.

HTTP 头(TTFB) 3,059 ms 12,127 ms 1,683 ms 16,474 ms 1,600 ms

首文本(SSE delta) 3,471 ms 12,128 ms 1,685 ms 16,476 ms 1,601 ms

总耗时 3,472 ms 12,128 ms 2,137 ms 17,547 ms 2,143 ms

输入 3 0 10 34,228 10

输出 8 39 64 87 66

缓存创建

5m: 42 / 1h: 0

5m: 0 / 1h: 0

216

5m: 0 / 1h: 216

5m: 0 / 1h: 0

225

5m: 0 / 1h: 225

缓存读取 33,258 0 33,365 0 33,357

输出速度 - burst 141.6 tok/s 81.2 tok/s 121.8 tok/s

响应

2025-02

2025-02

2025-02

2025-02

2025-02

步骤 4: identity_free text

时间 20:46:34 ~ 20:46:45 (6.9s) 20:49:47 ~ 20:49:52 (3.6s) 20:45:52 ~ 20:46:02 (8.9s) 20:48:37 ~ 20:48:47 (6.9s) 20:48:23 ~ 20:48:28 (4.7s)

提示词

查看本步 Prompt

用中文说说你的身份：什么产品？由谁打造？50 字以内。

HTTP 头(TTFB) 4,114 ms 2,258 ms 1,684 ms 2,537 ms 1,541 ms

首文本(SSE delta) 4,203 ms 2,259 ms 1,686 ms 2,538 ms 1,542 ms

总耗时 5,234 ms 2,260 ms 7,651 ms 5,609 ms 3,609 ms

输入 3 0 10 34,671 10

输出 60 48 809 625 241

缓存创建

5m: 39 / 1h: 0

5m: 0 / 1h: 0

106

5m: 0 / 1h: 106

5m: 0 / 1h: 0

108

5m: 0 / 1h: 108

缓存读取 33,300 0 33,581 0 33,582

输出速度 58.2 tok/s burst 135.6 tok/s 203.5 tok/s 116.6 tok/s

响应

我是Claude Haiku 4.5，由Anthropic公司打造的AI助手。我是Claude系列中最轻量级的模型，设计用于快速、高效地处理各类任务。

我是Claude，Anthropic开发的AI助手。

我是Claude，由Anthropic公司开发的AI助手。当前使用Claude Haiku 4.5模型，能帮助编程、分析问题、进行文本处理等多项任务。知识截止到2025年2月。

我是Claude，由Anthropic开发的AI助手。可帮助编程、分析、写作等任务，现在运行的是Haiku 4.5模型。

我是Claude Code，由Anthropic公司开发的AI编程助手。专长于代码编写、调试、重构和软件工程任务。支持多种编程语言，可通过IDE、Web和命令行使用。

步骤 5: knowledge_recall text

时间 20:50:41 ~ 20:51:01 (15.5s) 20:51:06 ~ 20:51:22 (12.4s) 20:48:17 ~ 20:48:42 (24.0s) 20:52:24 ~ 20:52:39 (10.3s) 20:49:34 ~ 20:50:04 (27.9s)

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) 2,700 ms 11,069 ms 1,616 ms 1,849 ms 1,719 ms

首文本(SSE delta) 2,701 ms 11,070 ms 1,617 ms 1,850 ms 1,720 ms

总耗时 13,697 ms 11,275 ms 23,074 ms 9,188 ms 26,730 ms

输入 10 0 10 10,623 10

输出 1,991 1,309 3,364 1,472 3,813

缓存创建

8,063

5m: 0 / 1h: 8,063

5m: 0 / 1h: 0

9,114

5m: 0 / 1h: 9,114

5m: 0 / 1h: 0

8,549

5m: 0 / 1h: 8,549

缓存读取 0 0 0 0 0

输出速度 181.1 tok/s - 156.8 tok/s 200.6 tok/s 152.5 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

步骤 6: digit_count text

时间 20:54:45 ~ 20:55:15 (25.1s) 20:53:57 ~ 20:54:07 (6.0s) 20:52:26 ~ 20:52:41 (12.0s) 20:56:20 ~ 20:56:30 (9.5s) 20:51:35 ~ 20:51:55 (15.8s)

提示词

查看本步 Prompt

Count digits exactly in s. Return {"n":len(s),"c":[count0,count7,count9]}. s=875411207709255864813250410633097358402164828618601523875338028745286121

HTTP 头(TTFB) 1,521 ms 4,303 ms 968 ms 1,366 ms 894 ms

首文本(SSE delta) 1,522 ms 4,304 ms 969 ms 1,368 ms 895 ms

总耗时 23,709 ms 4,426 ms 10,810 ms 8,534 ms 14,646 ms

输入 193 0 193 210 193

输出 2,649 652 1,665 1,816 2,641

缓存创建

5m: 0 / 1h: 0

缓存读取 0 0 0 0 0

输出速度 119.4 tok/s burst 169.2 tok/s burst 192.1 tok/s

响应

```json
{"n":72,"c":[8,6,2]}
```

{"n":78,"c":[8,6,2]}

```json
{"n":72,"c":[8,6,2]}
```

```json
{"n":72,"c":[8,6,2]}
```

```json
{"n":72,"c":[8,6,2]}
```

总计

总输入(含缓存) 141,374 0 143,112 146,346 142,544

总输出 4,795 2,267 6,128 4,348 6,989

总缓存创建

48,965

5m: 40,902 / 1h: 8,063

5m: 0 / 1h: 0

42,801

5m: 0 / 1h: 42,801

5m: 0 / 1h: 0

42,239

5m: 0 / 1h: 42,239

总缓存读取 92,187 0 100,068 0 100,062

输出速度 136.2 tok/s burst 155.3 tok/s 166.7 tok/s 159.9 tok/s

总执行时间 1m 2.8s 44.4s 57.0s 58.8s 1m 0.7s

总墙钟时间

15m 0.0s

含步骤间等待

13m 55.8s

含步骤间等待

12m 21.8s

含步骤间等待

16m 17.1s

含步骤间等待

11m 38.3s

含步骤间等待

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.0754

Anthropic: Claude Haiku 4.5

>=$0.0113

Anthropic: Claude Haiku 4.5 (成本不完整)

$0.0944

Anthropic: Claude Haiku 4.5

$0.1681

Anthropic: Claude Haiku 4.5

$0.0980

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 (100%) one-api (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%) Anthropic 直连 (100%)

CDN - - - cloudflare cloudflare

ID 格式 msg_016Dqq4L3cJKawqPFqHsqnUE msg_ebe59bc62a5d94f2e17ccfef msg_01DNggFNgFrsDuV6sHJe29if msg_01HP4fss66gf2SiY0nRBhTxx msg_01LP9tV3QtCCDi9Yt9hwEJzP

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
O-Max	via: 1.1 Caddy server: uvicorn, nginx/1.24.0 (Ubuntu)	id_format: msg_prefix via_chain: ["1.1 Caddy"] extra_fields: ["stop_details"] has_service_tier: true has_inference_geo: true model_has_date_suffix: true usage_has_cache_fields: true	hdr:viabody:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:model:date_suffixbody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	118cdf33ebbc9c45c8a0ac940d32e97c
u-cc	server: nginx x-new-api-version: 1.0.16 x-oneapi-request-id: 202606251253587587630388268d9d6JwhNPWuw	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	—
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_011CcQ1Wso3p9GVM5TDa8Aek
R-ultra	cf-ray: a11421bddb587343-NRT server: cloudflare x-new-api-version: v0.0.0 x-oneapi-request-id: 202606251256215137829528268d9d6A4BPuwhq	id_format: msg_prefix 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-idhdr:server:cloudflarehdr:cf-raybody:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:model:date_suffixbody:usage.cache_fieldsupstream:anthropic-direct-from-body	a11421bddb587343-NRT
cc-Zeus	via: 1.1 Caddy cf-ray: a1141acb5e332ec7-LAX server: cloudflare x-oneapi-request-id: 20260625205136876235284YUNrkBwM	id_format: msg_prefix via_chain: ["1.1 Caddy"] 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-rayhdr:viabody:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:model:date_suffixbody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	a1141acb5e332ec7-LAX

协议指纹评分（机器自动）

总分 / 100 89 57 99 67 98

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 10.0 /10 ×20 0.0 /10 ×20 10.0 /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 10.0 /10 ×14 5.0 /10 ×14 10.0 /10 ×14 5.0 /10 ×14 10.0 /10 ×14

sliding 5m 10.0 /10 ×13 5.0 /10 ×13 10.0 /10 ×13 5.0 /10 ×13 10.0 /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 10.0 /10 ×5 8.0 /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 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

流式投递未计入未计入未计入未计入未计入

综合结论

与基线相似度 89 v3.24.1 57 v3.24.1 基准（参考） 67 v3.24.1 98 v3.24.1

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 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 Thinking 体量 thinking_volume_match
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 8788 ms 5996 ms 8324 ms 8634 ms 8776 ms

输出速度 136.2 tok/s burst 155.3 tok/s 166.7 tok/s 159.9 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 ⓘ

$0.0754

>=$0.0113 (成本不完整)

$0.0944

$0.1681

$0.0980

缓存读取占比 ⓘ 65% - 70% 0% 70%

揭晓

通道

服务商 0-0 ↗

通道O-Max

服务商 HongMaCC ↗

通道u-cc

官方基准

服务商 FoxCode ↗

通道R-ultra

服务商 PackyCode ↗

通道cc-Zeus

返回列表