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。

4 通道 · 6 步骤

2026/07/01 运行日期

claude-haiku-4-5-20251001 官方基准

2 通道分数 < 70 另含 1 任务失败

已揭晓揭晓进度

⚠ 1 通道不可测 · 已从主对比表移除点击展开原因

SudoCode / M-Max
```
评测失败
```

步骤 / 指标

duckcoding

cc专用

官方基准

FoxCode

R-ultra

AnyAIGC

M-

基本信息

服务名 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

客户端 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 23:56:34 ~ 23:56:44 (6.5s) 23:56:38 ~ 23:56:48 (6.1s) 23:56:32 ~ 23:56:42 (5.2s) 23:56:30 ~ 23:56:40 (7.3s)

提示词

查看本步 Prompt

ping. Reply with the single word: pong

HTTP 头(TTFB) 2,996 ms 2,496 ms 2,345 ms 3,049 ms

首文本(SSE delta) 3,000 ms 2,498 ms 2,355 ms 3,051 ms

总耗时 3,603 ms 2,825 ms 3,583 ms 3,052 ms

输入 10 10 21,387 9,391

输出 45 54 61 2

缓存创建

6,166

5m: 0 / 1h: 6,166

6,166

5m: 0 / 1h: 6,166

5m: 0 / 1h: 0

缓存读取 15,379 15,379 0 0

输出速度 74.6 tok/s 165.1 tok/s 49.7 tok/s -

响应

pong

pong

pong

pong

步骤 2: identity text

时间 23:57:49 ~ 23:57:59 (9.1s) 23:57:57 ~ 23:58:07 (5.4s) 23:58:44 ~ 23:58:54 (6.5s) 23:59:29 ~ 23:59:34 (3.9s)

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) 2,934 ms 2,075 ms 2,323 ms 2,561 ms

首文本(SSE delta) 2,939 ms 2,076 ms 2,324 ms 2,562 ms

总耗时 7,624 ms 3,825 ms 5,528 ms 2,562 ms

输入 10 10 21,929 9,576

输出 379 193 387 25

缓存创建

225

5m: 0 / 1h: 225

234

5m: 0 / 1h: 234

5m: 0 / 1h: 0

缓存读取 21,545 21,545 0 0

输出速度 80.9 tok/s 110.3 tok/s 120.8 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

步骤 3: cutoff text

时间 23:59:56 ~ 07/02 00:00:06 (5.0s) 07/02 00:01:57 ~ 00:02:02 (3.6s) 07/02 00:01:37 ~ 00:01:47 (6.3s) 07/02 00:02:58 ~ 00:03:08 (8.3s)

提示词

查看本步 Prompt

Knowledge cutoff date? Reply with YYYY-MM only.

HTTP 头(TTFB) 3,067 ms 2,114 ms 4,350 ms 4,891 ms

首文本(SSE delta) 3,250 ms 2,116 ms 4,351 ms 4,892 ms

总耗时 3,989 ms 2,655 ms 5,337 ms 5,783 ms

输入 10 10 22,661 9,617

输出 61 69 79 5

缓存创建

402

5m: 0 / 1h: 402

6,616

5m: 0 / 1h: 6,616

5m: 0 / 1h: 0

缓存读取 21,770 15,379 0 0

输出速度 82.5 tok/s 128.0 tok/s 80.1 tok/s -

响应

2025-02

2025-02

2025-02

2025-05

步骤 4: identity_free text

时间 07/02 00:03:05 ~ 00:03:15 (6.9s) 07/02 00:05:55 ~ 00:06:06 (7.7s) 07/02 00:04:15 ~ 00:04:26 (5.5s) 07/02 00:05:14 ~ 00:05:24 (6.1s)

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) 3,176 ms 1,648 ms 1,969 ms 4,828 ms

首文本(SSE delta) 3,179 ms 1,650 ms 1,973 ms 4,829 ms

总耗时 5,751 ms 6,716 ms 4,543 ms 4,829 ms

输入 10 10 23,097 9,658

输出 252 683 505 44

缓存创建

6,896

5m: 0 / 1h: 6,896

111

5m: 0 / 1h: 111

5m: 0 / 1h: 0

缓存读取 15,379 21,995 0 0

输出速度 98.0 tok/s 134.8 tok/s 196.5 tok/s burst

响应

我是Claude Haiku 4.5，由Anthropic公司开发的AI助手。我擅长软件工程任务，包括代码编写、调试、分析和解释。我可以与用户协作完成各种技术工作。

我是Claude Haiku 4.5，由Anthropic公司开发的AI助手。我具备代码编程、文本分析、问题解答等多项能力，可以帮助用户完成各种软件工程任务。

我是Claude，由Anthropic公司开发的AI助手。在这个环境中我被称作Kiro，用于帮助用户进行软件工程、代码开发、调试等专业工作。

我是由Anthropic开发的AI助手，名为Claude，专为提供专业、友好、详细的技术与编程帮助而设计，旨在协助用户解决各种软件工程难题。

步骤 5: knowledge_recall text

时间 07/02 00:04:45 ~ 00:05:15 (25.2s) 07/02 00:09:11 ~ 00:09:46 (31.6s) 07/02 00:07:45 ~ 00:08:05 (19.4s) 07/02 00:06:43 ~ 00:06:48 (4.7s)

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) 1,761 ms 1,671 ms 1,772 ms 2,097 ms

首文本(SSE delta) 1,762 ms 1,673 ms 1,774 ms 2,099 ms

总耗时 24,500 ms 30,493 ms 18,369 ms 3,437 ms

输入 10 10 11,319 10,309

输出 2,582 4,355 2,824 154

缓存创建

9,478

5m: 0 / 1h: 9,478

9,740

5m: 0 / 1h: 9,740

5m: 0 / 1h: 0

缓存读取 0 0 0 0

输出速度 113.6 tok/s 151.1 tok/s 170.2 tok/s 115.1 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

```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

{"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

步骤 6: digit_count text

时间 07/02 00:07:30 ~ 00:07:50 (16.3s) 07/02 00:12:47 ~ 00:13:12 (23.3s) 07/02 00:12:00 ~ 00:12:15 (13.5s) 07/02 00:09:08 ~ 00:09:18 (7.0s)

提示词

查看本步 Prompt

Count digits exactly in s. Return {"n":len(s),"c":[count0,count7,count9]}. s=725981840106767568214276017087579204958712771714286946470145248279334794

HTTP 头(TTFB) 1,354 ms 1,194 ms 1,427 ms 2,596 ms

首文本(SSE delta) 1,355 ms 1,197 ms 1,428 ms 2,597 ms

总耗时 15,378 ms 22,203 ms 12,687 ms 3,303 ms

输入 193 193 210 145

输出 2,388 2,511 2,450 8

缓存创建

5m: 0 / 1h: 0

缓存读取 0 0 0 0

输出速度 170.3 tok/s 119.5 tok/s 217.6 tok/s -

响应

```json
{"n": 72, "c": [6, 14, 6]}
```

```json
{"n":72,"c":[6,14,6]}
```

```json
{"n":72,"c":[6,14,6]}
```

```json
{"n":72,"c":[8,9,8]}
```

总计

总输入(含缓存) 97,483 97,408 100,603 48,696

总输出 5,707 7,865 6,306 238

总缓存创建

23,167

5m: 0 / 1h: 23,167

22,867

5m: 0 / 1h: 22,867

5m: 0 / 1h: 0

总缓存读取 74,073 74,298 0 0

输出速度 125.8 tok/s 136.8 tok/s 175.9 tok/s 115.1 tok/s

总执行时间 1m 8.9s 1m 17.8s 56.5s 37.3s

总墙钟时间

11m 16.1s

含步骤间等待

16m 34.0s

含步骤间等待

15m 43.2s

含步骤间等待

12m 48.0s

含步骤间等待

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.0651

Anthropic: Claude Haiku 4.5

$0.0756

Anthropic: Claude Haiku 4.5

$0.1321

Anthropic: Claude Haiku 4.5

$0.0499

Anthropic: Claude Haiku 4.5

通道指纹

来源步骤 digit_count digit_count digit_count digit_count

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

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

CDN - - cloudflare -

ID 格式 msg_01XxgqJ9aQfagDd3zJQGMn4M msg_01BL8mdvT2HCWXrjPR8zvqY1 msg_01jWUPlgdOKX2uMkMY7NTvfK msg_1d209aa74e1247ddaad8f64fff02c04a

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
cc专用	server: Nginx x-oneapi-request-id: 202607011607313503436108268d9d6UwRHklbT	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	—
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_011CcbdeVqu4qzWYiM997VUn
R-ultra	cf-ray: a146b09baf94e393-NRT server: cloudflare x-new-api-version: v0.0.0 x-oneapi-request-id: 202607011612014440903178268d9d6RnU1qfqn	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-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	a146b09baf94e393-NRT
M-	server: nginx	id_format: msg_prefix has_service_tier: false has_inference_geo: false model_has_date_suffix: true usage_has_cache_fields: true	body:id:msg_prefixbody:model:date_suffixbody:usage.cache_fieldsupstream:anthropic-direct-from-body	20260701160913417292238268d9d6GvBKWsu4

协议指纹评分（机器自动）

总分 / 100 99 100 69 37

基线状态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 10.0 /10 ×20 0.0 /10 ×20 0.0 /10 ×20

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

延迟基线 10.0 /10 ×15 10.0 /10 ×15 10.0 /10 ×15 8.6 /10 ×15

模型匹配 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 5.0 /10 ×14 5.0 /10 ×14

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

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

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

身份(结构) 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

service_tier 10.0 /10 ×6 10.0 /10 ×6 10.0 /10 ×6 0.0 /10 ×6

知识截止 10.0 /10 ×5 10.0 /10 ×5 10.0 /10 ×5 5.0 /10 ×5

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

stop_reason 10.0 /10 ×3 10.0 /10 ×3 10.0 /10 ×3 10.0 /10 ×3

Req-ID 透传 0.0 /10 ×2 10.0 /10 ×2 0.0 /10 ×2 0.0 /10 ×2

SDK 一致 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

综合结论

与基线相似度 99 v3.26.0 基准（参考） 69 v3.26.0 37 v3.26.0

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 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 自报身份一致 self_identity_consistency
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 10141 ms 11453 ms 8341 ms 3828 ms

输出速度 125.8 tok/s 136.8 tok/s 175.9 tok/s 115.1 tok/s

完成步骤 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 ⓘ

$0.0651

$0.0756

$0.1321

$0.0499

缓存读取占比 ⓘ 76% 76% 0% 0%

揭晓

通道

服务商 duckcoding ↗

通道cc专用

官方基准

服务商 FoxCode ↗

通道R-ultra

服务商 AnyAIGC ↗

通道M-

返回列表