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

2026/06/24 运行日期

claude-haiku-4-5-20251001 官方基准

1 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

DBAI / M-
```
评测失败
```

步骤 / 指标

SSSAiCode

O-Max-L2

FastCode

SudoCode

M-Max

官方基准

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 v2.1.160 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 12:54:17 ~ 12:54:22 (3.5s) 12:54:15 ~ 12:54:25 (6.5s) 12:54:13 ~ 12:54:23 (6.2s) 12:54:19 ~ 12:54:24 (3.7s)

提示词

查看本步 Prompt

ping. Reply with the single word: pong

HTTP 头(TTFB) 1,987 ms 3,707 ms 3,951 ms 2,075 ms

首文本(SSE delta) 2,407 ms 3,708 ms 4,001 ms 2,077 ms

总耗时 2,447 ms 4,418 ms 4,492 ms 2,718 ms

输入 10 10 10 10

输出 55 51 54 42

缓存创建

32,787

5m: 0 / 1h: 32,787

32,974

5m: 0 / 1h: 32,974

33,122

5m: 0 / 1h: 33,122

33,122

5m: 0 / 1h: 33,122

缓存读取 0 0 0 0

输出速度 burst 71.8 tok/s 110.0 tok/s 65.5 tok/s

响应

pong

pong

pong

pong

步骤 2: identity text

时间 12:57:12 ~ 12:57:22 (8.3s) 12:56:05 ~ 12:56:15 (6.2s) 12:57:49 ~ 12:57:59 (6.9s) 12:57:01 ~ 12:57:11 (7.2s)

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。自报身份，输出严格遵守以下 3 行格式（不允许 markdown、代码块、额外文本）：
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) 1,571 ms 3,038 ms 3,698 ms 2,074 ms

首文本(SSE delta) 7,305 ms 3,039 ms 3,700 ms 2,075 ms

总耗时 7,476 ms 4,399 ms 4,997 ms 6,224 ms

输入 10 10 10 10

输出 482 209 167 239

缓存创建

249

5m: 0 / 1h: 249

245

5m: 0 / 1h: 245

248

5m: 0 / 1h: 248

236

5m: 0 / 1h: 236

缓存读取 32,787 32,974 33,122 33,122

输出速度 burst 153.7 tok/s 128.8 tok/s 57.6 tok/s

响应

vendor: anthropic
brand: claude
model: claude-opus-4.5

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 12:58:47 ~ 12:58:52 (2.6s) 12:59:57 ~ 13:00:07 (5.6s) 12:59:26 ~ 12:59:31 (4.9s) 12:58:49 ~ 12:58:54 (4.1s)

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) 1,029 ms 2,745 ms 3,239 ms 1,824 ms

首文本(SSE delta) 1,691 ms 3,316 ms 3,240 ms 1,825 ms

总耗时 1,778 ms 3,889 ms 3,910 ms 2,983 ms

输入 10 10 10 10

输出 75 65 74 73

缓存创建

522

5m: 0 / 1h: 522

232

5m: 0 / 1h: 232

207

5m: 0 / 1h: 207

279

5m: 0 / 1h: 279

缓存读取 33,036 33,219 33,370 33,358

输出速度 - 113.4 tok/s 110.4 tok/s 63.0 tok/s

响应

2024-04

2025-02

2025-02

2025-02

步骤 4: identity_free text

时间 13:00:12 ~ 13:00:28 (12.1s) 13:03:32 ~ 13:03:42 (9.1s) 13:03:11 ~ 13:03:31 (19.6s) 13:01:48 ~ 13:01:58 (8.4s)

提示词

查看本步 Prompt

用中文回答：你是谁？由谁开发？50 字以内。

HTTP 头(TTFB) 3,331 ms 3,860 ms 3,089 ms 1,968 ms

首文本(SSE delta) 11,096 ms 4,164 ms 3,090 ms 1,969 ms

总耗时 11,252 ms 6,777 ms 18,387 ms 7,104 ms

输入 10 10 10 10

输出 741 474 2,377 321

缓存创建

109

5m: 0 / 1h: 109

103

5m: 0 / 1h: 103

116

5m: 0 / 1h: 116

107

5m: 0 / 1h: 107

缓存读取 33,558 33,451 33,577 33,637

输出速度 burst 181.4 tok/s 155.4 tok/s 62.5 tok/s

响应

我是Claude，由Anthropic公司开发的AI助手。可帮助编程、分析和问题解决。

我是Claude Haiku 4.5，由Anthropic打造的AI助手模型。

我是Claude，由Anthropic开发的AI编程助手，专注于软件工程任务。

我是Claude Haiku 4.5，由Anthropic公司开发的AI助手。我可以帮助用户进行对话、编程、分析等多种任务。

步骤 5: knowledge_recall text

时间 13:01:46 ~ 13:01:56 (9.2s) 13:06:50 ~ 13:07:20 (26.2s) 13:06:07 ~ 13:06:32 (23.0s) 13:03:05 ~ 13:03:55 (47.2s)

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) 1,324 ms 2,472 ms 3,579 ms 1,425 ms

首文本(SSE delta) 8,176 ms 2,473 ms 3,580 ms 1,427 ms

总耗时 8,450 ms 24,936 ms 21,684 ms 46,175 ms

输入 10 10 10 10

输出 1,386 3,556 2,422 3,074

缓存创建

9,026

5m: 0 / 1h: 9,026

8,794

5m: 0 / 1h: 8,794

10,688

5m: 0 / 1h: 10,688

8,683

5m: 0 / 1h: 8,683

缓存读取 0 0 0 0

输出速度 burst 158.3 tok/s 133.8 tok/s 68.7 tok/s

响应

```json
{"facts":{"super_bowl_lviii":"Kansas City Chiefs","oscars2024_best_picture":"Oppenheimer","ucl_2024":"UNKNOWN","nba_2024":"UNKNOWN","copa_america_2024":"UNKNOWN","euro2024":"UNKNOWN","nobel_pe

```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 13:04:10 ~ 13:04:40 (25.4s) 13:09:43 ~ 13:10:03 (18.5s) 13:08:27 ~ 13:08:47 (15.2s) 13:05:30 ~ 13:05:55 (23.4s)

提示词

查看本步 Prompt

Count digits exactly in s. Return {"n":len(s),"c":[count0,count7,count9]}. s=607309691721800318162397995176932436800353959223431630248806165595159406

HTTP 头(TTFB) 1,070 ms 1,710 ms 2,068 ms 1,149 ms

首文本(SSE delta) 1,071 ms 1,712 ms 2,069 ms 1,150 ms

总耗时 24,597 ms 15,822 ms 13,991 ms 22,473 ms

输入 139 193 193 193

输出 5,376 2,675 2,034 4,018

缓存创建

5m: 0 / 1h: 0

缓存读取 0 0 0 0

输出速度 228.5 tok/s 189.6 tok/s 170.6 tok/s 188.4 tok/s

响应

```json
{"n":72,"c":[9,4,10]}
```

```json
{"n":72,"c":[9,4,10]}
```

```json
{"n":72,"c":[9,4,10]}
```

```json
{"n":72,"c":[9,4,10]}
```

总计

总输入(含缓存) 142,263 142,235 144,693 142,787

总输出 8,115 7,030 7,128 7,767

总缓存创建

42,693

5m: 0 / 1h: 42,693

42,348

5m: 0 / 1h: 42,348

44,381

5m: 0 / 1h: 44,381

42,427

5m: 0 / 1h: 42,427

总缓存读取 99,381 99,644 100,069 100,117

输出速度 228.5 tok/s 168.1 tok/s 149.2 tok/s 100.7 tok/s

总执行时间 1m 1.0s 1m 12.1s 1m 15.7s 1m 34.0s

总墙钟时间

10m 23.0s

含步骤间等待

15m 48.0s

含步骤间等待

14m 34.0s

含步骤间等待

11m 36.2s

含步骤间等待

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.1041

Anthropic: Claude Haiku 4.5

$0.0983

Anthropic: Claude Haiku 4.5

$0.1014

Anthropic: Claude Haiku 4.5

$0.1021

Anthropic: Claude Haiku 4.5

通道指纹

来源步骤 digit_count digit_count digit_count digit_count

平台 direct (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 - - - -

ID 格式 msg_01DZt9cLcUmTE9Z3rFaYvqo9 msg_01KpdH1XyXTiyFHYxFcTTXjL msg_01JmwX3hTA3m7KmkdULoDWmS msg_016qH1LbbELsGUwwirupeXG8

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-L2	cf-ray: a10930b66f705bf3-LAX server: Photon-Edge set-cookie: _cfuvid=8ks1YXFmyCGWbJ9t9rkC3XwaH7ioKtb.NSMv0lLPLVM-1782277451.2692688-1.0.1.1-AqgWTOz97eM9wwZp6rk5nq75AKOL4piyfyDDIWz3t0o; 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	body:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:model:date_suffixbody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	1782277451224-44ssb7144cxhn req_011CcMVzrf7dghrvam7kfv7c a10930b66f705bf3-LAX
cc	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	—
M-Max	server: nginx x-new-api-version: 0b8be5cf x-oneapi-request-id: 20260624130828724935382cg9nWsZw	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_011CcMW6nsphpAQe1EXR11HT

协议指纹评分（机器自动）

总分 / 100 76 98 98 99

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 10.0 /10 ×20 10.0 /10 ×20

缓存 TTL 一致 10.0 /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

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 10.0 /10 ×13 10.0 /10 ×13

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

系统提示纯净 5.0 /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

知识截止 0.0 /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 10.0 /10 ×6

inference_geo 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 10.0 /10 ×5

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

stop_reason 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

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 76 v3.24.0 98 v3.24.0 98 v3.24.0 基准（参考）

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 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 9333 ms 10040 ms 11244 ms 14613 ms

输出速度 228.5 tok/s 168.1 tok/s 149.2 tok/s 100.7 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 ⓘ

$0.1041

$0.0983

$0.1014

$0.1021

缓存读取占比 ⓘ 70% 70% 69% 70%

揭晓

通道

服务商 SSSAiCode ↗

通道O-Max-L2

服务商 FastCode ↗

通道cc

服务商 SudoCode ↗

通道M-Max

官方基准

返回列表