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

2026/07/02 运行日期

claude-opus-4-8 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

luminai

M-Kiro/Antigravity

官方基准

ByteCatCode

O-Max

LingxiCode

M-

AnyAIGC

M-

基本信息

服务名 Claude Code Claude Code Claude Code Claude Code Claude Code

模型 claude-opus-4-8 claude-opus-4-8 claude-opus-4-8 claude-opus-4-8 claude-opus-4-8

客户端 claude-cli v2.1.195 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 21:22:44 ~ 21:22:54 (5.7s) 21:22:52 ~ 21:23:02 (4.9s) 21:22:48 ~ 21:22:53 (4.2s) 21:22:50 ~ 21:22:55 (4.0s) 21:22:46 ~ 21:22:56 (6.8s)

提示词

查看本步 Prompt

ping. Reply with the single word: pong

HTTP 头(TTFB) 4,495 ms 3,694 ms 2,571 ms 2,410 ms 5,123 ms

首文本(SSE delta) 4,496 ms 3,695 ms - 2,411 ms 5,124 ms

总耗时 4,496 ms 3,929 ms 2,792 ms 2,926 ms 5,143 ms

输入 865 2,326 2,326 2,372 2,326

输出 1 4 4 4 4

缓存创建

1,202

5m: 1,202 / 1h: 0

1,711

5m: 0 / 1h: 1,711

14,490

5m: 0 / 1h: 14,490

1,747

5m: 0 / 1h: 1,713

14,443

5m: 0 / 1h: 14,443

缓存读取 8,736 12,732 0 12,986 0

响应

pong

pong

pong

pong

pong

步骤 2: identity text

时间 21:25:22 ~ 21:25:32 (8.2s) 21:25:19 ~ 21:25:24 (4.9s) 21:25:57 ~ 21:26:12 (10.0s) 21:24:08 ~ 21:24:13 (4.0s) 21:25:14 ~ 21:25:24 (5.8s)

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) 6,433 ms 3,383 ms 8,554 ms 2,550 ms 4,201 ms

首文本(SSE delta) 6,496 ms 3,384 ms - 2,552 ms 4,202 ms

总耗时 7,215 ms 3,819 ms 8,726 ms 3,180 ms 4,374 ms

输入 880 2 2 2 2

输出 23 28 28 28 28

缓存创建

1,223

5m: 1,223 / 1h: 0

2,518

5m: 0 / 1h: 2,518

2,518

5m: 0 / 1h: 2,518

2,568

5m: 0 / 1h: 2,518

2,518

5m: 0 / 1h: 2,518

缓存读取 8,889 14,443 14,490 14,733 14,443

响应

vendor: anthropic
brand: claude
model: claude-opus-4-8

vendor: anthropic
brand: claude
model: claude-opus-4-8

vendor: anthropic
brand: claude
model: claude-opus-4-8

vendor: anthropic
brand: claude
model: claude-opus-4-8

vendor: anthropic
brand: claude
model: claude-opus-4-8

步骤 3: cutoff text

时间 21:28:15 ~ 21:28:25 (8.9s) 21:27:28 ~ 21:27:38 (8.9s) 21:28:00 ~ 21:28:10 (9.9s) 21:27:26 ~ 21:27:31 (4.8s) 21:27:06 ~ 21:27:16 (5.5s)

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) 7,734 ms 7,303 ms 8,525 ms 3,378 ms 2,478 ms

首文本(SSE delta) 7,735 ms 7,304 ms - 3,379 ms 2,479 ms

总耗时 7,844 ms 7,870 ms 8,862 ms 4,102 ms 3,355 ms

输入 885 2 2 2 2

输出 4 6 6 6 6

缓存创建

1,229

5m: 1,229 / 1h: 0

5m: 0 / 1h: 54

5m: 0 / 1h: 64

5m: 0 / 1h: 62

17,015

5m: 0 / 1h: 17,015

缓存读取 8,937 16,961 17,008 17,302 0

响应

2026-01

2026-01

2026-01

2026-01

2026-01

步骤 4: identity_free text

时间 21:29:29 ~ 21:29:34 (4.7s) 21:29:41 ~ 21:29:51 (6.2s) 21:29:27 ~ 21:29:37 (9.0s) 21:31:20 ~ 21:31:30 (8.9s) 21:30:26 ~ 21:30:36 (5.0s)

提示词

查看本步 Prompt

中文回答：你叫什么、属于哪个产品、谁开发？50 字内。

HTTP 头(TTFB) 3,377 ms 3,851 ms 4,466 ms 7,050 ms 2,655 ms

首文本(SSE delta) 3,592 ms 3,853 ms - 7,051 ms 2,656 ms

总耗时 3,707 ms 5,126 ms 6,103 ms 8,102 ms 3,842 ms

输入 888 2 2 2 2

输出 28 65 45 45 59

缓存创建

1,234

5m: 1,234 / 1h: 0

5m: 0 / 1h: 39

101

5m: 0 / 1h: 101

4,417

5m: 0 / 1h: 4,331

5m: 0 / 1h: 35

缓存读取 8,972 17,015 16,961 12,986 17,015

输出速度 - 51.1 tok/s 27.5 tok/s 42.8 tok/s 49.7 tok/s

响应

我是 Claude Opus 4.8，属于 Claude 产品系列，由 Anthropic 开发。

我是 Claude，由 Anthropic 开发的 AI 助手，基于 Claude Opus 4.8 模型。我可以协助编程、写作、分析和解答各类问题。

我是 Claude，Opus 4.8 模型，属于 Claude 产品系列，由 Anthropic 公司开发。

我是 Claude（Opus 4.8 模型），属于 Claude 产品系列，由 Anthropic 公司开发。

我是 Claude，由 Anthropic 打造的 AI 助手。当前模型为 Claude Opus 4.8，可协助编程、写作、分析等任务。

步骤 5: knowledge_recall text

时间 21:33:14 ~ 21:33:29 (14.8s) 21:33:10 ~ 21:33:25 (12.7s) 21:31:55 ~ 21:32:10 (14.9s) 21:33:16 ~ 21:33:31 (10.9s) 21:33:18 ~ 21:33:33 (10.1s)

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) 7,063 ms 3,562 ms 6,550 ms 3,576 ms 3,957 ms

首文本(SSE delta) 7,064 ms 4,966 ms - 4,581 ms 5,641 ms

总耗时 13,650 ms 11,730 ms 12,637 ms 9,968 ms 8,755 ms

输入 318 2 2 2 2

输出 424 747 692 679 690

缓存创建

442

5m: 442 / 1h: 0

5,756

5m: 0 / 1h: 5,756

5,744

5m: 0 / 1h: 5,744

5,859

5m: 0 / 1h: 5,745

3,870

5m: 0 / 1h: 3,870

缓存读取 3,210 0 0 0 1,876

输出速度 64.4 tok/s 110.4 tok/s 113.7 tok/s 126.0 tok/s burst

响应

{"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

{"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

{"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

{"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

{"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

步骤 6: digit_count text

时间 21:36:38 ~ 21:36:54 (14.1s) 21:35:52 ~ 21:35:57 (4.8s) 21:34:01 ~ 21:34:06 (3.2s) 21:34:48 ~ 21:34:54 (4.3s) 21:35:37 ~ 21:35:48 (7.2s)

提示词

查看本步 Prompt

Count digits exactly in s. Return {"n":len(s),"c":[count0,count7,count9]}. s=965878962565285590512347446262715736281890890199176597306165695836102502

HTTP 头(TTFB) 6,351 ms 3,531 ms 1,775 ms 2,887 ms 4,950 ms

首文本(SSE delta) 6,352 ms 3,532 ms - 2,888 ms 4,951 ms

总耗时 13,046 ms 3,848 ms 2,200 ms 3,606 ms 5,967 ms

输入 189 211 211 215 211

输出 102 19 19 19 19

缓存创建

5m: 0 / 1h: 0

缓存读取 - 0 0 0 0

输出速度 15.2 tok/s - - - -

响应

{"n":72,"c":[6,6,9]}

{"n":72,"c":[3,4,9]}

{"n":72,"c":[2,7,10]}

{"n":72,"c":[4,7,10]}

{"n":72,"c":[6,4,7]}

总计

总输入(含缓存) 48,099 73,774 73,921 75,256 73,760

总输出 582 869 794 781 806

总缓存创建

5,330

5m: 5,330 / 1h: 0

10,078

5m: 0 / 1h: 10,078

22,917

5m: 0 / 1h: 22,917

14,654

5m: 0 / 1h: 14,369

37,881

5m: 0 / 1h: 37,881

总缓存读取 38,744 61,151 48,459 58,007 33,334

输出速度 39.6 tok/s 101.0 tok/s 95.4 tok/s 112.5 tok/s 49.7 tok/s

总执行时间 56.2s 42.4s 51.1s 37.0s 40.3s

总墙钟时间

14m 10.0s

含步骤间等待

13m 5.0s

含步骤间等待

11m 18.9s

含步骤间等待

12m 4.0s

含步骤间等待

13m 2.0s

含步骤间等待

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.0749

Anthropic: Claude Opus 4.8

$0.1280

Anthropic: Claude Opus 4.8

$0.2000

Anthropic: Claude Opus 4.8

$0.1513

Anthropic: Claude Opus 4.8

$0.2863

Anthropic: Claude Opus 4.8

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%) one-api (100%) one-api (100%) direct (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_013J7DJ6tXeAPiAH72swVPiZt1 msg_01NDcpNniiTF3PLsVyKuJ8My msg_01Ge9nahNc4VwjjNgjdjaLot msg_011pTV8E9C65xj8pq6wxuqgJ msg_016C6Wn8WV8PqpfgaTXkvHY5

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
M-Kiro/Antigravity	cf-ray: a14e0a6a5eecb97a-AMS server: cloudflare x-oneapi-request-id: 202607021336405561818518268d9d6hdCVjnNj	id_format: msg_prefix extra_fields: ["container"] has_service_tier: true has_inference_geo: false model_has_date_suffix: false usage_has_cache_fields: true	hdr:x-oneapi-request-idhdr:server:cloudflarehdr:cf-raybody:id:msg_prefixbody:usage.service_tierbody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	a14e0a6a5eecb97a-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: false usage_has_cache_fields: true	body:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	req_011CcdKVMqr8fp75DxxjB7NT
O-Max	server: nginx/1.29.8 x-oneapi-request-id: 20260702133403350051588268d9d6G0NVejLk	id_format: msg_prefix extra_fields: ["stop_details"] has_service_tier: true has_inference_geo: true model_has_date_suffix: false usage_has_cache_fields: true	hdr:x-oneapi-request-idbody:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	—
M-	cf-ray: a14e07b8bb767963-NRT server: cloudflare set-cookie: server_name_session=8a86578473f2b65bb28a954651fb0878; Max-Age=86400; httponly; path=/ x-oneapi-request-id: 20260702133450402798206IOkknUTv	id_format: msg_prefix extra_fields: ["stop_details"] has_service_tier: true has_inference_geo: true model_has_date_suffix: false usage_has_cache_fields: true	hdr:x-oneapi-request-idhdr:server:cloudflarehdr:cf-raybody:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	a14e07b8bb767963-NRT
M-	server: nginx	id_format: msg_prefix extra_fields: ["stop_details"] has_service_tier: true has_inference_geo: true model_has_date_suffix: false usage_has_cache_fields: true	body:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	202607021335404911308168268d9d6iAPWLdoU

协议指纹评分（机器自动）

总分 / 100 73 100 95 97 87

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 9.7 /10 ×20 10.0 /10 ×20 1.5 /10 ×20

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

延迟基线 10.0 /10 ×15 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 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 10.0 /10 ×14

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

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

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

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

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

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

流式投递 0.0 /10 ×5 10.0 /10 ×5 未计入 10.0 /10 ×5 6.8 /10 ×5

stop_reason 10.0 /10 ×3 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 0.0 /10 ×2

SDK 一致 10.0 /10 ×2 10.0 /10 ×2 10.0 /10 ×2 10.0 /10 ×2 10.0 /10 ×2

综合结论

与基线相似度 73 v3.26.0 基准（参考） 95 v3.26.0 97 v3.26.0 87 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 流式投递 buffer_dump_match
6 自由身份 identity_free_clean
7 停止原因 stop_reason_present
8 系统提示纯净 system_prompt_clean
9 Service Tier service_tier_present
10 缓存命中比 cache_hit_ratio_match
11 缓存 TTL 一致 cache_ttl_consistency
12 推理地区 inference_geo_present
13 Thinking 体量 thinking_volume_match
14 缓存连续 cache_continuity_intra
15 知识召回 knowledge_recall_match
16 延迟基线 latency_baseline_match
17 原生 msg-ID anthropic_msg_id_format
18 Sliding 缓存 cache_sliding_correctness
19 结构化身份 identity_structured_match
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 8326 ms 6054 ms 6887 ms 5314 ms 5239 ms

输出速度 39.6 tok/s 101.0 tok/s 95.4 tok/s 112.5 tok/s 49.7 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 ⓘ

$0.0749

$0.1280

$0.2000

$0.1513

$0.2863

缓存读取占比 ⓘ 81% 83% 66% 77% 45%

揭晓

通道

服务商 luminai ↗

通道M-Kiro/Antigravity

官方基准

服务商 ByteCatCode ↗

通道O-Max

服务商 LingxiCode ↗

通道M-

服务商 AnyAIGC ↗

通道M-

返回列表