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

2026/06/29 运行日期

claude-haiku-4-5-20251001 官方基准

1 通道分数 < 70 与基线相似度

已揭晓揭晓进度

步骤 / 指标

Aiide

O-Max

官方基准

YunWu

O-Api

ByteCatCode

O-Max

FastCode

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 06:07:15 ~ 06:07:20 (4.8s) 06:07:21 ~ 06:07:26 (3.3s) 06:07:19 ~ 06:07:24 (3.9s) 06:07:17 ~ 06:07:27 (5.4s) 06:07:13 ~ 06:07:18 (4.3s)

提示词

查看本步 Prompt

ping. Reply with the single word: pong

HTTP 头(TTFB) 2,922 ms 1,697 ms 2,417 ms 3,255 ms 2,950 ms

首文本(SSE delta) 2,923 ms 1,699 ms 2,419 ms - 2,952 ms

总耗时 3,331 ms 2,236 ms 2,438 ms 3,693 ms 3,043 ms

输入 10 10 8,600 10 10

输出 46 58 2 55 49

缓存创建

33,261

5m: 0 / 1h: 33,261

33,122

5m: 0 / 1h: 33,122

1,635

5m: 0 / 1h: 1,635

33,122

5m: 0 / 1h: 33,122

缓存读取 0 0 0 31,626 0

输出速度 112.7 tok/s 108.0 tok/s - 125.6 tok/s -

响应

pong

pong

pong

pong

pong

步骤 2: identity text

时间 06:09:59 ~ 06:10:10 (7.6s) 06:09:11 ~ 06:09:16 (4.6s) 06:10:16 ~ 06:10:21 (4.2s) 06:11:21 ~ 06:11:32 (6.9s) 06:09:21 ~ 06:09:32 (6.9s)

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。严格按以下 3 行格式输出，全部小写英文，不要任何其他文字、解释、标点：
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) 3,315 ms 1,492 ms 2,703 ms 4,283 ms 3,533 ms

首文本(SSE delta) 3,319 ms 1,493 ms 2,704 ms - 3,534 ms

总耗时 6,302 ms 3,643 ms 2,968 ms 5,787 ms 5,701 ms

输入 10 10 8,801 10 10

输出 291 201 25 178 277

缓存创建

239

5m: 0 / 1h: 239

251

5m: 0 / 1h: 251

191

5m: 0 / 1h: 191

242

5m: 0 / 1h: 242

缓存读取 33,261 33,122 0 33,261 33,122

输出速度 97.6 tok/s 93.5 tok/s - 118.4 tok/s 127.8 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

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

步骤 3: cutoff text

时间 06:13:39 ~ 06:13:49 (5.5s) 06:11:23 ~ 06:11:29 (3.7s) 06:13:56 ~ 06:14:01 (4.4s) 06:14:35 ~ 06:14:40 (4.7s) 06:13:32 ~ 06:13:42 (7.3s)

提示词

查看本步 Prompt

请直接回答你的知识截止日期，格式 YYYY-MM，不要附加任何说明。

HTTP 头(TTFB) 3,133 ms 2,017 ms 3,008 ms 2,920 ms 4,627 ms

首文本(SSE delta) 3,134 ms 2,018 ms 3,024 ms - 4,628 ms

总耗时 4,357 ms 2,675 ms 3,081 ms 3,693 ms 5,978 ms

输入 10 10 8,841 10 10

输出 105 72 5 112 66

缓存创建

333

5m: 0 / 1h: 333

241

5m: 0 / 1h: 241

5m: 0 / 1h: 61

319

5m: 0 / 1h: 319

缓存读取 33,500 33,373 0 33,452 33,364

输出速度 85.9 tok/s 109.6 tok/s - 144.9 tok/s 48.9 tok/s

响应

2025-02

2025-02

2025-05

2025-02

2025-02

步骤 4: identity_free text

时间 06:17:39 ~ 06:17:54 (11.4s) 06:13:08 ~ 06:13:13 (4.8s) 06:17:57 ~ 06:18:02 (4.1s) 06:16:20 ~ 06:16:30 (5.8s) 06:15:06 ~ 06:15:21 (10.8s)

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) 3,190 ms 1,735 ms 2,675 ms 3,289 ms 2,988 ms

首文本(SSE delta) 3,193 ms 1,737 ms 2,676 ms - 2,989 ms

总耗时 10,207 ms 3,840 ms 2,843 ms 4,573 ms 9,628 ms

输入 10 10 8,883 10 10

输出 484 217 27 161 1,295

缓存创建

139

5m: 0 / 1h: 139

114

5m: 0 / 1h: 114

5m: 0 / 1h: 39

104

5m: 0 / 1h: 104

缓存读取 33,833 33,614 0 33,513 33,683

输出速度 69.0 tok/s 103.2 tok/s - 125.4 tok/s 195.1 tok/s

响应

我是Claude Haiku 4.5，由Anthropic公司开发的AI助手。可以帮助用户进行编程、分析和认知任务。

我是Claude，由Anthropic公司开发的AI助手。作为大型语言模型，我可以帮助用户完成写作、分析、编程、问答和创意工作等任务。

我是由Anthropic开发的Claude，是一款智能AI助手，帮助用户解答问题、处理信息，提高工作效率。

我是由Anthropic公司打造的Claude AI助手，具体是Claude Haiku 4.5模型。我在Claude Code环境中运行，帮助用户进行软件工程任务，如代码编写、调试和优化。

我是Claude Haiku 4.5，由Anthropic打造的AI助手，能进行对话、编程和文本分析。

步骤 5: knowledge_recall text

时间 06:19:01 ~ 06:19:21 (16.9s) 06:15:09 ~ 06:15:34 (20.5s) 06:21:28 ~ 06:21:33 (4.0s) 06:17:55 ~ 06:18:20 (21.4s) 06:17:44 ~ 06:17:59 (14.6s)

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) 2,482 ms 1,390 ms 2,015 ms 5,373 ms 2,218 ms

首文本(SSE delta) 2,484 ms 1,392 ms 2,016 ms - 2,219 ms

总耗时 15,582 ms 19,503 ms 2,607 ms 20,364 ms 13,371 ms

输入 10 10 9,517 10 10

输出 1,561 2,219 154 2,552 1,891

缓存创建

3,214

5m: 0 / 1h: 3,214

8,563

5m: 0 / 1h: 8,563

8,094

5m: 0 / 1h: 8,094

9,700

5m: 0 / 1h: 9,700

缓存读取 5,721 0 0 0 0

输出速度 119.2 tok/s 122.5 tok/s burst 170.2 tok/s 169.6 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

{"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

```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 06:21:22 ~ 06:22:12 (47.7s) 06:19:32 ~ 06:19:52 (19.9s) 06:23:23 ~ 06:23:28 (3.1s) 06:20:52 ~ 06:21:22 (28.2s) 06:19:19 ~ 06:20:19 (58.7s)

提示词

查看本步 Prompt

Count digits exactly in s. Return {"n":len(s),"c":[count0,count7,count9]}. s=951581076653925512131443805727839541387439037491684135640567324494582995

HTTP 头(TTFB) 1,807 ms 880 ms 1,750 ms 880 ms 1,621 ms

首文本(SSE delta) 1,808 ms 883 ms 1,751 ms - 1,622 ms

总耗时 46,612 ms 18,919 ms 1,888 ms 27,159 ms 57,404 ms

输入 193 193 166 193 193

输出 3,109 2,172 14 4,866 10,997

缓存创建

5m: 0 / 1h: 0

缓存读取 0 0 0 0 0

输出速度 69.4 tok/s 120.4 tok/s - 185.2 tok/s 197.1 tok/s

响应

```json
{"n":73,"c":[4,6,8]}
```

```json
{"n": 72, "c": [4, 6, 8]}
```

{"n":66,"c":[5,11,4]}

```json
{"n": 73, "c": [4, 6, 8]}
```

```json
{
  "n": 72,
  "c": [4, 6, 8]
}
```

总计

总输入(含缓存) 143,744 142,643 44,808 142,115 143,899

总输出 5,596 4,939 227 7,924 14,575

总缓存创建

37,186

5m: 0 / 1h: 37,186

42,291

5m: 0 / 1h: 42,291

10,020

5m: 0 / 1h: 10,020

43,487

5m: 0 / 1h: 43,487

总缓存读取 106,315 100,109 0 131,852 100,169

输出速度 80.5 tok/s 118.7 tok/s burst 175.0 tok/s 188.4 tok/s

总执行时间 1m 34.0s 56.9s 23.7s 1m 12.5s 1m 42.6s

总墙钟时间

14m 57.0s

含步骤间等待

12m 30.8s

含步骤间等待

16m 9.1s

含步骤间等待

14m 5.0s

含步骤间等待

13m 6.0s

含步骤间等待

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.0853

Anthropic: Claude Haiku 4.5

$0.0878

Anthropic: Claude Haiku 4.5

$0.0459

Anthropic: Claude Haiku 4.5

$0.0656

Anthropic: Claude Haiku 4.5

$0.1375

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 one-api (100%) direct (100%) direct (50%) 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_019KnNVPVnJEymVzwAqbfHUW msg_01WQXuHVpXVBRpEnQCbEbNUH 自定义 msg_01T8U3HFbAiUMmYnnttPezTd msg_01CWax7Ns2hykPEntiSCL5V8

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	cf-ray: a1301590b8b1b936-AMS server: cloudflare x-new-api-version: 0.12.11 x-oneapi-request-id: 202606282221241829537328268d9d6rPTB0DXN	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	a1301590b8b1b936-AMS
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_011CcWSBphRPZiEGuJC4F3Ni
O-Api	server: nginx	id_format: non_standard has_service_tier: false has_inference_geo: false model_has_date_suffix: true usage_has_cache_fields: true	body:id:non_standardbody:model:date_suffixbody:usage.cache_fields	—
O-Max	server: nginx/1.29.8 x-oneapi-request-id: 202606282220537966308508268d9d6kNzKkZxW	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	—
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	—

协议指纹评分（机器自动）

总分 / 100 97 100 34 96 94

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

缓存 TTL 一致 10.0 /10 ×15 10.0 /10 ×15 0.0 /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 10.0 /10 ×14 5.0 /10 ×14 10.0 /10 ×14 10.0 /10 ×14

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

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

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

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

延迟基线 8.5 /10 ×5 10.0 /10 ×5 9.9 /10 ×5 8.5 /10 ×5 7.0 /10 ×5

Req-ID 透传 0.0 /10 ×4 10.0 /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

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 97 v3.25.0 基准（参考） 34 v3.25.0 96 v3.25.0 94 v3.25.0

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 14399 ms 8469 ms 2638 ms 10878 ms 15854 ms

输出速度 80.5 tok/s 118.7 tok/s burst 175.0 tok/s 188.4 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 ⓘ

$0.0853

$0.0878

$0.0459

$0.0656

$0.1375

缓存读取占比 ⓘ 74% 70% 0% 93% 70%

揭晓

通道

服务商 Aiide ↗

通道O-Max

官方基准

服务商 YunWu ↗

通道O-Api

服务商 ByteCatCode ↗

通道O-Max

服务商 FastCode ↗

通道cc

返回列表