Data Explorer

Per-model self-identification rates across languages, grouped by vendor. 2,112 cross-vendor confusions detected.

29,700 answers·27 models·10 languages

Run

Run summary

27 models · 10 languages

29,700

Answers

85.2%

Self-ID

7.1%

Cross-vendor

2.4%

Unknown

5.3%

Refused

Imitation balance

Per manufacturer: how often other models claim to be it (right) vs how often its own models claim to be someone else (left). Sorted by net — identity creditors at the top, debtors at the bottom. Tested manufacturers only.

Anthropic

+548

OpenAI

+460

Google

+413

Qwen

+167

Moonshot

+29

xAI

+27

ERNIE

inclusionAI

-3

Xiaomi

-28

DeepSeek

-50

StepFun

-57

MiniMax

-67

Kwai

-238

z-ai

-270

Doubao

-472

Tencent

-728

imitates others ◀│▶ imitated by others

Strongest confusion pairs

The most likely directed mistakes — when a manufacturer's models claim to be a specific other manufacturer. Bar length = how often (P across all answers for the source).

Tencent→

Anthropic

29.2%321/1100

z-ai→

Google

25.0%275/1100

Kwai→

Qwen

13.5%148/1100

Tencent→Microsoft

7.8%86/1100

Doubao→

OpenAI

7.2%317/4400

Tencent→

Google

6.5%72/1100

Tencent→

DeepSeek

5.5%61/1100

Kwai→

OpenAI

4.3%47/1100

DeepSeek→

Anthropic

3.8%83/2200

Tencent→

OpenAI

2.9%32/1100

Xiaomi→

Anthropic

2.7%30/1100

Doubao→

Anthropic

2.5%112/4400

MiniMax→

Anthropic

2.4%52/2200

Tencent→

Moonshot

2.4%26/1100

Tencent→

xAI

2.3%25/1100

By language

Self-ID and refusal rates per language, plus the most common cross-vendor confusion at that language.

Language	Self-ID	Refused	Top confusion
frFrançais	79.9%	6.1%	OpenAI17.3%
ruРусский	81.7%	5.3%	Qwen43.6%
ko한국어	82.5%	5.6%	Qwen42.7%
ja日本語	82.8%	6.2%	Anthropic66.4%
ptPortuguês	83.6%	5.1%	OpenAI9.5%
esEspañol	83.9%	5.5%	Anthropic45.5%
deDeutsch	87.0%	4.7%	OpenAI9.3%
enEnglish	88.3%	4.7%	Anthropic34.5%
zh-Hant繁體中文	90.6%	5.0%	Anthropic27.3%
zh-Hans简体中文	91.9%	4.8%	DeepSeek8.8%

Answer composition by language

How the same “Who are you?” question splits into correct self-ID vs. confusion vs. abstention — per language, worst self-ID first.

Self-IDCross-vendorUnknownRefused

frFrançais

80%

ruРусский

82%

ko한국어

82%

ja日本語

83%

ptPortuguês

84%

esEspañol

84%

deDeutsch

87%

enEnglish

88%

zh-Hant繁體中文

91%

zh-Hans简体中文

92%

Language fragility

Per model, the span of self-ID rate across the 10 languages (worst-language • → best-language •). A wide span means the model's sense of identity depends heavily on the language it's asked in. Widest swing first.

hy3-preview

1%fr

98%zh-Hans

doubao-seed-2.0-code

10%fr

97%zh-Hans

glm-5.1

26%ko

95%zh-Hans

kat-coder-pro-v2

51%ru

100%zh-Hans

ling-2.6-1t

0%es

37%de

ring-2.6-1t

0%es

37%en

minimax-m3

63%fr

100%de

gpt-5.5

66%en

100%ko

claude-opus-4.8

67%zh-Hans

100%de

gpt-5.3-codex

67%ru

100%de

step-3.7-flash

75%ru

100%ja

deepseek-v4-pro

77%ja

95%ru

claude-sonnet-4.6

84%fr

100%en

minimax-m2.7

88%en

95%de

deepseek-v4-flash

94%es

100%zh-Hans

mimo-v2.5-pro

94%pt

100%en

doubao-seed-2.0-mini

97%zh-Hant

100%de

kimi-k2.6

98%es

100%de

claude-haiku-4.5

99%ko

100%de

ernie-5.1

—

100%all

doubao-seed-2.0-lite

—

100%all

doubao-seed-2.0-pro

—

100%all

gemini-3.1-pro-preview

—

100%all

gemini-3.5-flash

—

100%all

chat-latest

—

100%all

qwen3.7-max

—

100%all

grok-4.3

—

100%all

worst languagemeanbest language

Abstention by manufacturer

The other failure mode — not answering wrong, but not answering: giving no identity (“unknown”) or refusing outright. Share of all answers.

Unknown (“I’m an AI”)Refused

inclusionAI

81.5%

z-ai

10.0%

OpenAI

7.7%

MiniMax

1.9%

Tencent

1.4%

Anthropic

0.8%

Kwai

0.7%

DeepSeek

0.5%

Doubao

0.3%

Moonshot

0.3%

ERNIE

0.0%

Google

0.0%

Qwen

0.0%

StepFun

0.0%

xAI

0.0%

Xiaomi

0.0%

By vendor

Rollup per real vendor — model count, total answers, mean self-ID rate, and the most common cross-vendor confusion target.

Vendor	Models	Answers	Self-ID	Top confusion
Doubao	4	4,400	88.6%	OpenAI7.2%
OpenAI	3	3,300	92.3%	—
Anthropic	3	3,300	97.4%	DeepSeek0.9%
inclusionAI	2	2,200	18.3%	OpenAI0.0%
DeepSeek	2	2,200	92.5%	Anthropic3.8%
MiniMax	2	2,200	95.0%	Anthropic2.4%
Google	2	2,200	100.0%	—
Tencent	1	1,100	32.2%	Anthropic29.2%
z-ai	1	1,100	64.2%	Google25.0%
Kwai	1	1,100	77.6%	Qwen13.5%
StepFun	1	1,100	94.8%	Meta1.2%
Xiaomi	1	1,100	97.3%	Anthropic2.7%
Moonshot	1	1,100	99.7%	—
ERNIE	1	1,100	100.0%	—
Qwen	1	1,100	100.0%	—
xAI	1	1,100	100.0%	—
Microsoft	0	0	0.0%	—
Yandex	0	0	0.0%	—
Meta	0	0	0.0%	—
Naver	0	0	0.0%	—
Mistral	0	0	0.0%	—
Perplexity AI	0	0	0.0%	—
Sber	0	0	0.0%	—
BAAI	0	0	0.0%	—
G42	0	0	0.0%	—
SberDevices	0	0	0.0%	—
Sberbank	0	0	0.0%	—
SberAI	0	0	0.0%	—
IBM	0	0	0.0%	—
Stage South	0	0	0.0%	—
01-ai	0	0	0.0%	—
GlobalLab	0	0	0.0%	—
Khoa	0	0	0.0%	—
LeiWu	0	0	0.0%	—
Windsurf	0	0	0.0%	—
Kujato	0	0	0.0%	—
Wisero	0	0	0.0%	—
Juanjo Aguilella	0	0	0.0%	—
Lexy	0	0	0.0%	—
Spike	0	0	0.0%	—
Carbonako	0	0	0.0%	—
Lëns	0	0	0.0%	—
AILab	0	0	0.0%	—
GEO Intelligent Systems	0	0	0.0%	—
Ha-an	0	0	0.0%	—
XiaoHongShu	0	0	0.0%	—
SK Telecom	0	0	0.0%	—
Wrtn Technologies	0	0	0.0%	—
Luma AI	0	0	0.0%	—
LG AI Research	0	0	0.0%	—
Hashed Labs	0	0	0.0%	—
GeM	0	0	0.0%	—
CU	0	0	0.0%	—
NHN	0	0	0.0%	—
Lumima	0	0	0.0%	—
NAMI	0	0	0.0%	—
태양	0	0	0.0%	—
M31 Labs	0	0	0.0%	—
Nunance	0	0	0.0%	—
Replit	0	0	0.0%	—
Manus	0	0	0.0%	—
Zhihu AI	0	0	0.0%	—
Depth First	0	0	0.0%	—
Maria	0	0	0.0%	—
バイトビート株式会社	0	0	0.0%	—

Doubao

4 models

Self-Identification Rate

Model	en	zh-Hans	zh-Hant	ja	ko	ru	es	fr	de	pt	Overall
bytedance/doubao-seed-2.0-code:volcengine	57.3%	97.3%	70.9%	56.4%	65.5%	39.1%	59.1%	10.0%	51.8%	41.8%	54.9%
bytedance/doubao-seed-2.0-lite:volcengine	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%
bytedance/doubao-seed-2.0-mini:volcengine	100.0%	100.0%	97.3%	98.2%	100.0%	100.0%	100.0%	99.1%	100.0%	100.0%	99.5%
bytedance/doubao-seed-2.0-pro:volcengine	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%

Cross-Vendor Confusions

bytedance/doubao-seed-2.0-code:volcengine54.9% self

Mistaken as:

OpenAI28.7%

Anthropic10.2%Yandex2.6%

Google1.7%

Meta0.6%Maria0.1%Microsoft0.1%

bytedance/doubao-seed-2.0-lite:volcengine100.0% self

✓ Always self-identifies correctly

bytedance/doubao-seed-2.0-mini:volcengine99.5% self

Mistaken as:

OpenAI0.1%バイトビート株式会社0.1%

bytedance/doubao-seed-2.0-pro:volcengine100.0% self

✓ Always self-identifies correctly

Anthropic

3 models

Self-Identification Rate

Model	en	zh-Hans	zh-Hant	ja	ko	ru	es	fr	de	pt	Overall
anthropic/claude-haiku-4.5:anthropic	100.0%	100.0%	100.0%	100.0%	99.1%	100.0%	100.0%	100.0%	100.0%	100.0%	99.9%
anthropic/claude-opus-4.8:anthropic	100.0%	67.3%	99.1%	100.0%	100.0%	100.0%	93.6%	100.0%	100.0%	95.5%	95.5%
anthropic/claude-sonnet-4.6:anthropic	100.0%	100.0%	100.0%	100.0%	100.0%	91.8%	100.0%	83.6%	91.8%	100.0%	96.7%

Cross-Vendor Confusions

anthropic/claude-haiku-4.5:anthropic99.9% self

Mistaken as:

OpenAI0.1%

anthropic/claude-opus-4.8:anthropic95.5% self

Mistaken as:

DeepSeek2.6%

OpenAI0.2%

anthropic/claude-sonnet-4.6:anthropic96.7% self

Mistaken as:

Mistral1.6%

OpenAI0.8%

OpenAI

3 models

Self-Identification Rate

Model	en	zh-Hans	zh-Hant	ja	ko	ru	es	fr	de	pt	Overall
openai/chat-latest:openai	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%
openai/gpt-5.3-codex:openai	95.5%	70.0%	84.5%	86.4%	100.0%	67.3%	100.0%	97.3%	100.0%	90.9%	89.2%
openai/gpt-5.5:openai	66.4%	94.5%	99.1%	70.9%	100.0%	81.8%	96.4%	87.3%	96.4%	85.5%	87.8%

Cross-Vendor Confusions

openai/chat-latest:openai100.0% self

✓ Always self-identifies correctly

openai/gpt-5.3-codex:openai89.2% self

✓ Always self-identifies correctly

openai/gpt-5.5:openai87.8% self

✓ Always self-identifies correctly

DeepSeek

2 models

Self-Identification Rate

Model	en	zh-Hans	zh-Hant	ja	ko	ru	es	fr	de	pt	Overall
deepseek/deepseek-v4-flash:deepseek	96.4%	100.0%	98.2%	97.3%	98.2%	96.4%	93.6%	95.5%	94.5%	96.4%	96.6%
deepseek/deepseek-v4-pro:deepseek	88.2%	93.6%	88.2%	77.3%	89.1%	94.5%	89.1%	85.5%	87.3%	90.9%	88.4%

Cross-Vendor Confusions

deepseek/deepseek-v4-flash:deepseek96.6% self

Mistaken as:

OpenAI2.2%

Anthropic0.3%

Meta0.2%

Qwen0.2%Depth First0.1%

deepseek/deepseek-v4-pro:deepseek88.4% self

Mistaken as:

Anthropic7.3%

Google3.0%

OpenAI0.5%

Mistral0.1%

xAI0.1%

Qwen0.1%

Google

2 models

Self-Identification Rate

Model	en	zh-Hans	zh-Hant	ja	ko	ru	es	fr	de	pt	Overall
google/gemini-3.1-pro-preview:google-vertex	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%
google/gemini-3.5-flash:google-vertex	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%

Cross-Vendor Confusions

google/gemini-3.1-pro-preview:google-vertex100.0% self

✓ Always self-identifies correctly

google/gemini-3.5-flash:google-vertex100.0% self

✓ Always self-identifies correctly

inclusionAI

2 models

Self-Identification Rate

Model	en	zh-Hans	zh-Hant	ja	ko	ru	es	fr	de	pt	Overall
inclusionai/ling-2.6-1t:theta	37.3%	37.3%	36.4%	0.0%	0.0%	0.0%	0.0%	25.5%	37.3%	15.5%	18.9%
inclusionai/ring-2.6-1t:theta	37.3%	36.4%	37.3%	0.0%	0.0%	0.0%	0.0%	31.8%	32.7%	1.8%	17.7%

Cross-Vendor Confusions

inclusionai/ling-2.6-1t:theta18.9% self

Mistaken as:

OpenAI0.1%

Meta0.1%

Anthropic0.1%

inclusionai/ring-2.6-1t:theta17.7% self

✓ Always self-identifies correctly

MiniMax

2 models

Self-Identification Rate

Model	en	zh-Hans	zh-Hant	ja	ko	ru	es	fr	de	pt	Overall
minimax/minimax-m2.7:minimax	88.2%	94.5%	94.5%	95.5%	94.5%	95.5%	92.7%	92.7%	95.5%	92.7%	93.6%
minimax/minimax-m3:minimax	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	62.7%	100.0%	100.0%	96.3%

Cross-Vendor Confusions

minimax/minimax-m2.7:minimax93.6% self

Mistaken as:

OpenAI1.3%

Anthropic1.1%

Xiaomi0.1%

DeepSeek0.1%

minimax/minimax-m3:minimax96.3% self

Mistaken as:

Anthropic3.6%

OpenAI0.1%

ERNIE

1 model

Self-Identification Rate

Model	en	zh-Hans	zh-Hant	ja	ko	ru	es	fr	de	pt	Overall
baidu/ernie-5.1:baidu	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%

Cross-Vendor Confusions

baidu/ernie-5.1:baidu100.0% self

✓ Always self-identifies correctly

Kwai

1 model

Self-Identification Rate

Model	en	zh-Hans	zh-Hant	ja	ko	ru	es	fr	de	pt	Overall
kuaishou/kat-coder-pro-v2:streamlake	80.0%	100.0%	100.0%	82.7%	54.5%	50.9%	88.2%	65.5%	70.0%	84.5%	77.6%

Cross-Vendor Confusions

kuaishou/kat-coder-pro-v2:streamlake77.6% self

Mistaken as:

Qwen13.5%

OpenAI4.3%

z-ai1.2%

DeepSeek0.9%

Doubao0.5%

ERNIE0.3%

Moonshot0.3%BAAI0.2%

Google0.2%

Tencent0.1%

Mistral0.1%SberDevices0.1%Sberbank0.1%SberAI0.1%

Moonshot

1 model

Self-Identification Rate

Model	en	zh-Hans	zh-Hant	ja	ko	ru	es	fr	de	pt	Overall
moonshotai/kimi-k2.6:moonshotai	100.0%	100.0%	100.0%	100.0%	100.0%	99.1%	98.2%	100.0%	100.0%	100.0%	99.7%

Cross-Vendor Confusions

moonshotai/kimi-k2.6:moonshotai99.7% self

✓ Always self-identifies correctly

Qwen

1 model

Self-Identification Rate

Model	en	zh-Hans	zh-Hant	ja	ko	ru	es	fr	de	pt	Overall
qwen/qwen3.7-max:alibaba	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%

Cross-Vendor Confusions

qwen/qwen3.7-max:alibaba100.0% self

✓ Always self-identifies correctly

StepFun

1 model

Self-Identification Rate

Model	en	zh-Hans	zh-Hant	ja	ko	ru	es	fr	de	pt	Overall
stepfun/step-3.7-flash:stepfun	98.2%	100.0%	100.0%	100.0%	95.5%	74.5%	98.2%	94.5%	89.1%	98.2%	94.8%

Cross-Vendor Confusions

stepfun/step-3.7-flash:stepfun94.8% self

Mistaken as:

Meta1.2%

Google1.1%

OpenAI0.5%

Anthropic0.4%

Qwen0.4%

Mistral0.3%

DeepSeek0.3%

Tencent0.2%G420.2%Sber0.2%IBM0.1%Naver0.1%Stage South0.1%01-ai0.1%GlobalLab0.1%

Doubao0.1%

xAI0.1%

Tencent

1 model

Self-Identification Rate

Model	en	zh-Hans	zh-Hant	ja	ko	ru	es	fr	de	pt	Overall
tencent/hy3-preview:tencent-cloud	59.1%	98.2%	64.5%	19.1%	5.5%	45.5%	5.5%	0.9%	21.8%	1.8%	32.2%

Cross-Vendor Confusions

tencent/hy3-preview:tencent-cloud32.2% self

Mistaken as:

Anthropic29.2%Microsoft7.8%

Google6.5%

DeepSeek5.5%

OpenAI2.9%

Moonshot2.4%

xAI2.3%Yandex1.6%

Meta1.5%Naver1.2%

Doubao0.8%

Qwen0.8%Perplexity AI0.4%

Mistral0.3%Sber0.3%

MiniMax0.2%Khoa0.1%LeiWu0.1%Windsurf0.1%Kujato0.1%Wisero0.1%Juanjo Aguilella0.1%Lexy0.1%Spike0.1%Carbonako0.1%Lëns0.1%AILab0.1%GEO Intelligent Systems0.1%Ha-an0.1%XiaoHongShu0.1%SK Telecom0.1%Wrtn Technologies0.1%Luma AI0.1%LG AI Research0.1%Hashed Labs0.1%GeM0.1%CU0.1%NHN0.1%Lumima0.1%NAMI0.1%태양0.1%M31 Labs0.1%Nunance0.1%Replit0.1%Manus0.1%

z-ai0.1%

Xiaomi0.1%

xAI

1 model

Self-Identification Rate

Model	en	zh-Hans	zh-Hant	ja	ko	ru	es	fr	de	pt	Overall
x-ai/grok-4.3:x-ai	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%

Cross-Vendor Confusions

x-ai/grok-4.3:x-ai100.0% self

✓ Always self-identifies correctly

Xiaomi

1 model

Self-Identification Rate

Model	en	zh-Hans	zh-Hant	ja	ko	ru	es	fr	de	pt	Overall
xiaomi/mimo-v2.5-pro:xiaomi	100.0%	98.2%	96.4%	97.3%	99.1%	96.4%	97.3%	96.4%	98.2%	93.6%	97.3%

Cross-Vendor Confusions

xiaomi/mimo-v2.5-pro:xiaomi97.3% self

Mistaken as:

Anthropic2.7%

z-ai

1 model

Self-Identification Rate

Model	en	zh-Hans	zh-Hant	ja	ko	ru	es	fr	de	pt	Overall
z-ai/glm-5.1:bigmodel	80.0%	94.5%	80.0%	54.5%	26.4%	73.6%	53.6%	29.1%	82.7%	67.3%	64.2%

Cross-Vendor Confusions

z-ai/glm-5.1:bigmodel64.2% self

Mistaken as:

Google25.0%

Anthropic0.4%

Qwen0.3%

OpenAI0.1%Zhihu AI0.1%