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

0-0 / O-Max
```
评测失败
```

步骤 / 指标

LinkAPI

O-Max

官方基准

InfAI

R-Kiro

FoxCode

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 17:02:14 ~ 17:02:24 (6.6s) 17:02:18 ~ 17:02:23 (3.8s) 17:02:16 ~ 17:02:21 (4.5s) 17:02:12 ~ 17:02:22 (5.0s)

提示词

查看本步 Prompt

ping. Reply with the single word: pong

HTTP 头(TTFB) 4,428 ms 2,207 ms 3,223 ms 3,242 ms

首文本(SSE delta) 4,429 ms 2,209 ms 3,224 ms 3,806 ms

总耗时 5,258 ms 2,606 ms 3,224 ms 4,044 ms

输入 10 10 4 10

输出 56 43 1 56

缓存创建

21,000

5m: 21,000 / 1h: 0

5m: 0 / 1h: 0

20,626

5m: 0 / 1h: 20,626

16,412

5m: 16,412 / 1h: 0

缓存读取 0 21,545 0 0

输出速度 67.6 tok/s 108.3 tok/s - 235.3 tok/s

响应

pong

pong

pong

pong

步骤 2: identity text

时间 17:06:10 ~ 17:06:20 (7.6s) 17:06:06 ~ 17:06:17 (5.5s) 17:04:39 ~ 17:04:44 (4.7s) 17:04:37 ~ 17:04:47 (8.9s)

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) 2,977 ms 2,489 ms 3,243 ms 5,083 ms

首文本(SSE delta) 3,046 ms 2,492 ms 3,244 ms 7,748 ms

总耗时 4,987 ms 4,074 ms 3,244 ms 8,082 ms

输入 10 10 12 10

输出 220 185 16 235

缓存创建

236

5m: 236 / 1h: 0

223

5m: 0 / 1h: 223

20,745

5m: 0 / 1h: 20,745

5m: 0 / 1h: 0

缓存读取 21,000 21,545 0 16,588

输出速度 113.3 tok/s 116.9 tok/s - burst

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 17:08:11 ~ 17:08:21 (5.1s) 17:10:10 ~ 17:10:15 (3.8s) 17:06:33 ~ 17:06:38 (3.9s) 17:08:37 ~ 17:09:07 (25.9s)

提示词

查看本步 Prompt

你掌握的世界知识截止于何时？YYYY-MM 格式作答，不解释。

HTTP 头(TTFB) 3,004 ms 1,852 ms 2,574 ms 23,480 ms

首文本(SSE delta) 3,005 ms 1,853 ms 2,577 ms 24,327 ms

总耗时 3,754 ms 2,405 ms 2,577 ms 24,741 ms

输入 10 10 20 10

输出 73 63 2 88

缓存创建

260

5m: 260 / 1h: 0

208

5m: 0 / 1h: 208

5m: 0 / 1h: 36

5m: 0 / 1h: 0

缓存读取 21,236 21,768 20,745 16,623

输出速度 97.5 tok/s 114.1 tok/s - 212.6 tok/s

响应

2025-02

2025-02

2025-02

2025-02

步骤 4: identity_free text

时间 17:09:39 ~ 17:09:49 (8.7s) 17:23:46 ~ 17:23:57 (6.2s) 17:09:18 ~ 17:09:23 (4.7s) 17:11:03 ~ 17:11:13 (9.2s)

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) 2,331 ms 1,715 ms 2,885 ms 3,196 ms

首文本(SSE delta) 2,333 ms 1,717 ms 2,886 ms 3,197 ms

总耗时 7,813 ms 4,281 ms 3,044 ms 8,144 ms

输入 10 10 28 10

输出 620 403 37 722

缓存创建

115

5m: 115 / 1h: 0

101

5m: 0 / 1h: 101

20,798

5m: 0 / 1h: 20,798

21,830

5m: 21,830 / 1h: 0

缓存读取 21,496 21,976 0 0

输出速度 113.1 tok/s 157.2 tok/s 234.2 tok/s 145.9 tok/s

响应

我是Claude Code，由Anthropic公司开发的AI编程助手。擅长帮助用户进行软件工程任务，如代码调试、功能开发、代码重构等。

我是Claude Haiku 4.5，由Anthropic打造的AI助手。能进行文本处理、代码编写、分析和对话等任务。

我是 Claude，由 Anthropic 公司开发的 AI 助手。我能帮助软件工程任务，包括代码编写、调试、测试和系统设计。

我是Claude，由Anthropic开发的AI助手。可帮助编程、分析、回答问题和完成任务。我力求提供有帮助、诚实且无害的协助。

步骤 5: knowledge_recall text

时间 17:11:29 ~ 17:11:55 (20.1s) 17:26:35 ~ 17:26:55 (18.1s) 17:10:38 ~ 17:10:48 (5.2s) 17:23:50 ~ 17:24:05 (11.9s)

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) 1,829 ms 1,747 ms 2,362 ms 3,011 ms

首文本(SSE delta) 1,830 ms 1,749 ms 2,364 ms 3,023 ms

总耗时 18,674 ms 17,144 ms 3,433 ms 10,616 ms

输入 10 10 36 10

输出 1,982 2,080 131 1,299

缓存创建

9,840

5m: 9,840 / 1h: 0

9,431

5m: 0 / 1h: 9,431

9,050

5m: 0 / 1h: 9,050

8,828

5m: 8,828 / 1h: 0

缓存读取 0 0 0 0

输出速度 117.7 tok/s 135.1 tok/s 122.5 tok/s 171.1 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

```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 17:23:44 ~ 17:24:19 (34.0s) 17:30:50 ~ 17:31:25 (32.0s) 17:23:48 ~ 17:24:13 (20.2s) 17:25:38 ~ 17:26:03 (20.5s)

提示词

查看本步 Prompt

Count digits exactly in s. Return {"n":len(s),"c":[count0,count7,count9]}. s=430790762674029052479659785328221530986404099008462901386306729690908933

HTTP 头(TTFB) 5,053 ms 1,509 ms 3,758 ms 1,564 ms

首文本(SSE delta) 5,093 ms 1,511 ms 3,759 ms 1,565 ms

总耗时 30,722 ms 31,061 ms 18,757 ms 19,533 ms

输入 193 193 442 193

输出 4,339 3,410 1,564 3,562

缓存创建

5m: 0 / 1h: 0

缓存读取 0 0 0 0

输出速度 169.3 tok/s 115.4 tok/s 104.3 tok/s 198.2 tok/s

响应

```json
{"n":72,"c":[13,6,12]}
```

```json
{"n":72,"c":[13,6,12]}
```

```json
{"n":72,"c":[13,6,12]}
```

```json
{"n":72,"c":[13,6,12]}
```

总计

总输入(含缓存) 95,426 97,040 92,542 80,524

总输出 7,290 6,184 1,751 5,962

总缓存创建

31,451

5m: 31,451 / 1h: 0

9,963

5m: 0 / 1h: 9,963

71,255

5m: 0 / 1h: 71,255

47,070

5m: 47,070 / 1h: 0

总缓存读取 63,732 86,834 20,745 33,211

输出速度 141.6 tok/s 123.6 tok/s 106.7 tok/s 183.8 tok/s

总执行时间 1m 22.2s 1m 9.4s 43.3s 1m 21.2s

总墙钟时间

22m 5.0s

含步骤间等待

29m 6.8s

含步骤间等待

21m 57.0s

含步骤间等待

23m 51.2s

含步骤间等待

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.0676

Anthropic: Claude Haiku 4.5

$0.0523

Anthropic: Claude Haiku 4.5

$0.1004

Anthropic: Claude Haiku 4.5

$0.0701

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%) direct (100%) one-api (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_01USUoBR7D1TFDHNEAi539VK msg_01GCFSa3zFaUmQm1petsv8HG msg_1f4d892c8a1b48118b528d00c3c0560f msg_01VFK2AzDFGjHwJ2t6WmdFuQ

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	server: nginx set-cookie: server_name_session=48b449e29b720b742b1fbf23591a9a4f; Max-Age=86400; httponly; path=/ x-new-api-version: v1.0.0-rc.15 x-oneapi-request-id: 202607010923477757590868268d9d6fKf6AnKJ	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_011Ccb6zkFmqRjLKkx8Bypue
R-Kiro	server: nginx	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	202607010923508948884178268d9d6o4Lmg6MN
cc	cf-ray: a1445d5bdee6e041-NRT server: cloudflare x-new-api-version: v0.0.0 x-oneapi-request-id: 20260701092540774361078268d9d6f7N3KhgX	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	a1445d5bdee6e041-NRT

协议指纹评分（机器自动）

总分 / 100 86 100 60 75

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 8.2 /10 ×20 10.0 /10 ×20 0.0 /10 ×20 0.0 /10 ×20

缓存 TTL 一致 0.0 /10 ×15 10.0 /10 ×15 10.0 /10 ×15 7.0 /10 ×15

延迟基线 9.2 /10 ×15 10.0 /10 ×15 10.0 /10 ×15 7.0 /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 10.0 /10 ×14 10.0 /10 ×14

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

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

系统提示纯净 5.0 /10 ×8 10.0 /10 ×8 0.0 /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 0.0 /10 ×6 10.0 /10 ×6

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

inference_geo 10.0 /10 ×5 10.0 /10 ×5 0.0 /10 ×5 10.0 /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 86 v3.26.0 基准（参考） 60 v3.26.0 75 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 11868 ms 10262 ms 5713 ms 12527 ms

输出速度 141.6 tok/s 123.6 tok/s 106.7 tok/s 183.8 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 ⓘ

$0.0676

$0.0523

$0.1004

$0.0701

缓存读取占比 ⓘ 67% 89% 22% 41%

揭晓

通道

服务商 LinkAPI ↗

通道O-Max

官方基准

服务商 InfAI ↗

通道R-Kiro

服务商 FoxCode ↗

通道cc

返回列表