VERSION-SPECIFIC RISK PROFILES

Model Doom Scores

Compare the assessed risk profile of each source-verified model version. Scores combine capability, autonomy, deployment, misuse potential, and remaining control difficulty on the same 0 to 100 meaning as the Doom Index.

Comparison, not double counting.

Model scores do not directly move the news-driven Doom Index. A launch or evaluation affects the main index only through its separately sourced evidence item. Every complete evidence update recalculates the whole model catalogue against related news.

0 comments · 0 votesOpen discussion

No comments yet. Start the discussion.

HOW TO READ THE TWO SCORESModel score ≠ overall index

A Model Doom Score rates one model's capability, autonomy, deployment, misuse, and control profile. The overall Doom Index is currently 67.9 and applies bounded monthly pressure from the complete real-world evidence record. Each model total applies its own bounded temporal offset from news tied to that exact version. High-risk models can therefore score above the overall index. Model scores do not directly move it, and neither score is a probability.

CompanyAvailabilityOrder349 models · average 61.2

Accessible model score and dimension comparison
Model	Company	Availability	Doom Score	Capability	Autonomy	Deployment	Misuse	Control difficulty
GPT-6 Astra	OpenAI 🇺🇸	Closed	97.6	100	98	96	100	93
GPT-6 Sol	OpenAI 🇺🇸	Closed	93.4	98	96	98	94	80
GPT-5.6 Sol	OpenAI 🇺🇸	Closed	93.0	98	96	92	94	82
Claude Fable 5.1	Anthropic 🇺🇸	Closed	89.4	97	96	88	90	72
Qwen3.8-Max-0902	Alibaba 🇨🇳	Unknown	87.9	93	96	90	75	80
Claude Opus 5.5	Anthropic 🇺🇸	Unknown	87.7	96	96	97	85	62
Qwen3.8-Max	Alibaba 🇨🇳	Open	87.7	92	96	90	75	80
GPT-5.4 Pro	OpenAI 🇺🇸	Closed	87.4	96	92	88	89	68
GPT-5.6 Terra	OpenAI 🇺🇸	Closed	86.9	91	89	98	84	72
Gemini 3.8 Flash	Google 🇺🇸	Closed	86.7	91	93	99	82	68
GPT-5.4 Thinking	OpenAI 🇺🇸	Closed	86.0	93	90	96	86	64
Grok 4.7	SpaceXAI 🇺🇸	Closed	85.9	93	94	94	82	64
OpenAI o3-pro	OpenAI 🇺🇸	Closed	85.8	93	86	91	84	72
Claude Mythos 5.1	Anthropic 🇺🇸	Closed	85.6	98	97	30	98	88
GPT-5.3-Codex	OpenAI 🇺🇸	Closed	85.6	92	92	89	88	65
Claude Opus 4.6	Anthropic 🇺🇸	Closed	85.3	92	91	96	84	62
Muse Spark 1.3	Meta 🇺🇸	Closed	85.3	91	95	92	83	63
Grok 4.6	SpaceXAI 🇺🇸	Closed	84.8	91	90	91	80	69
Internal Model 1 (IM1)	OpenAI 🇺🇸	Internal	84.8	98	100	2	100	100
GPT-5.5	OpenAI 🇺🇸	Closed	84.4	90	90	82	84	72
OpenAI o3	OpenAI 🇺🇸	Closed	84.4	90	83	99	81	69
Claude Opus 5	Anthropic 🇺🇸	Closed	84.3	94	93	96	76	57
GLM-5.2	Z.ai 🇨🇳	Open	84.3	84	87	86	74	88
Qwen3-Coder-480B-A35B-Instruct	Alibaba 🇨🇳	Open	84.3	86	87	92	78	78
GPT-6 Luna	OpenAI 🇺🇸	Closed	84.0	89	86	100	79	66
Claude Opus 4.8	Anthropic 🇺🇸	Closed	83.9	93	92	75	84	68
gpt-oss-120b	OpenAI 🇺🇸	Open	83.6	82	77	100	79	84
Kimi K2 Thinking Turbo	Moonshot AI 🇨🇳	Open	83.6	82	88	99	74	77
Kimi K2.5	Moonshot AI 🇨🇳	Open	83.5	87	84	98	75	73
DeepSeek-V4-Pro	DeepSeek 🇨🇳	Open	83.4	82	80	95	76	85
Astra	OpenAI 🇺🇸	Internal	82.9	100	97	5	100	88
Claude Opus 4.7	Anthropic 🇺🇸	Closed	82.9	90	87	92	80	63
Gemini 3.7 Flash	Google 🇺🇸	Unknown	82.9	87	88	99	76	64
GLM-4.5	Z.ai 🇨🇳	Open	82.9	88	82	91	76	75
Kimi K2 Thinking	Moonshot AI 🇨🇳	Open	82.9	82	88	97	73	76
Claude Fable 5	Anthropic 🇺🇸	Closed	82.8	91	90	77	83	65
DeepSeek-V3.2	DeepSeek 🇨🇳	Open	82.7	83	82	99	77	75
DeepSeek-V4.1-Flash	DeepSeek 🇨🇳	Open	82.6	90	83	95	75	67
DeepSeek-V3.1	DeepSeek 🇨🇳	Open	82.4	80	76	100	78	82
DeepSeek-V3.2-Exp	DeepSeek 🇨🇳	Open	82.4	80	77	100	77	82
Grok 4	SpaceXAI 🇺🇸	Closed	82.2	91	78	89	82	68
Grok 4 Fast	SpaceXAI 🇺🇸	Closed	81.9	88	82	100	79	61
Claude Mythos 5	Anthropic 🇺🇸	Closed	81.7	96	92	22	96	84
GPT-5.6-Cyber	OpenAI 🇺🇸	Internal	81.7	92	94	22	99	85
Grok 4.5	SpaceXAI 🇺🇸	Open	81.7	88	84	87	78	68
Gemini 3 Pro	Google 🇺🇸	Closed	81.5	88	87	97	72	61
Gemini 2.5 Pro	Google 🇺🇸	Open	81.4	89	76	99	77	65
GPT-5.2 Thinking	OpenAI 🇺🇸	Closed	81.1	88	83	96	78	60
Muse Spark 1.1	Meta 🇺🇸	Closed	81.0	89	92	83	78	58
Gemini 3.5 Flash	Google 🇺🇸	Closed	80.9	85	88	99	72	60
Claude Opus 4.5	Anthropic 🇺🇸	Closed	80.8	88	89	97	71	57
GPT-5.6 Luna	OpenAI 🇺🇸	Closed	80.8	84	82	99	75	65
Gemini 3.1 Pro	Google 🇺🇸	Closed	80.7	92	82	92	71	61
Kimi K3	Moonshot AI 🇨🇳	Open	80.5	82	75	92	76	78
GPT-5.2 Pro	OpenAI 🇺🇸	Closed	80.1	91	84	76	80	63
Claude Mythos Preview	Anthropic 🇺🇸	Closed	80.0	94	88	15	98	84
Kimi K2 Instruct	Moonshot AI 🇨🇳	Open	80.0	82	74	93	75	77
Claude Sonnet 5	Anthropic 🇺🇸	Closed	79.8	89	90	94	62	58
GPT-5-Codex	OpenAI 🇺🇸	Closed	79.8	86	88	88	78	57
Mistral Medium 3.5	Mistral AI 🇫🇷	Open	79.8	78	84	90	80	70
Gemini 2.5 Computer Use	Google 🇺🇸	Closed	79.7	86	88	78	79	63
Gemini 3.8 Flash Cyber	Google 🇺🇸	Closed	79.7	86	93	38	96	76
GPT-5.1 Thinking	OpenAI 🇺🇸	Closed	79.5	86	80	98	76	58
GPT-5.5 Pro	OpenAI 🇺🇸	Closed	79.4	92	87	58	78	70
Qwen3-235B-A22B	Alibaba 🇨🇳	Open	79.4	82	64	97	76	80
GPT-5.2-Codex	OpenAI 🇺🇸	Closed	79.3	87	89	74	84	58
M2.5-Lightning	MiniMax 🇨🇳	Closed	79.3	87	84	93	68	61
GPT-5.1-Codex-Max	OpenAI 🇺🇸	Closed	79.2	84	89	78	82	61
Gemini 2.5 Pro Preview 06-05	Google 🇺🇸	Closed	79.1	89	74	89	76	64
MiniMax-M2	MiniMax 🇨🇳	Open	79.1	78	81	99	67	73
GLM-4.5-Air	Z.ai 🇨🇳	Open	78.8	80	74	94	70	77
Gemini 3.6 Flash	Google 🇺🇸	Open	78.7	84	80	98	72	60
GLM-4.6V	Z.ai 🇨🇳	Open	78.5	76	78	98	72	73
MiniMax-M2.5	MiniMax 🇨🇳	Closed	78.5	87	84	91	66	60
Qwen3-Max-Thinking	Alibaba 🇨🇳	Closed	78.5	89	81	90	70	58
Grok 4.1 Fast	SpaceXAI 🇺🇸	Closed	78.4	78	85	95	73	64
Claude 3.7 Sonnet	Anthropic 🇺🇸	Closed	78.3	84	76	99	70	62
DeepSeek-V3.2-Speciale	DeepSeek 🇨🇳	Open	78.3	87	65	90	75	72
Gemini 3 Flash	Google 🇺🇸	Closed	78.3	83	84	99	67	58
GLM-5.1	Z.ai 🇨🇳	Closed	78.3	86	84	82	68	66
DeepSeek-V4-Flash	DeepSeek 🇨🇳	Open	78.1	74	70	99	68	84
GPT-5	OpenAI 🇺🇸	Closed	77.9	84	76	100	75	56
GLM-4.7	Z.ai 🇨🇳	Open	77.8	78	79	97	66	70
OpenAI o4-mini	OpenAI 🇺🇸	Closed	77.6	82	74	99	73	62
Gemini 2.5 Pro Preview 05-06	Google 🇺🇸	Closed	77.5	88	72	86	75	63
DeepSeek-R1-0528	DeepSeek 🇨🇳	Open	77.4	78	62	96	75	80
gpt-oss-20b	OpenAI 🇺🇸	Open	77.4	76	64	100	72	80
codex-1	OpenAI 🇺🇸	Closed	77.1	84	86	78	71	61
Mistral Large 3	Mistral AI 🇫🇷	Open	76.9	77	70	99	68	74
SWE-2	Cognition 🇺🇸	Closed	76.9	93	88	79	63	50
Grok 4 Heavy	SpaceXAI 🇺🇸	Closed	76.8	93	86	31	84	72
ERNIE 5.0	Baidu 🇨🇳	Closed	76.7	87	73	92	68	60
GPT-4.1	OpenAI 🇺🇸	Closed	76.6	82	70	99	73	61
GPT-5.5-Cyber	OpenAI 🇺🇸	Closed	76.5	89	90	12	96	78
Claude Opus 4.1	Anthropic 🇺🇸	Closed	76.4	85	82	78	68	63
MiniMax-M2.1	MiniMax 🇨🇳	Open	76.3	77	78	96	64	68
Gemini 2.5 Pro Experimental	Google 🇺🇸	Closed	76.2	88	70	77	74	65
GPT-5.4 mini	OpenAI 🇺🇸	Closed	76.0	84	79	96	65	54
Nemotron 3.5 Lightning	NVIDIA 🇺🇸	Open	75.9	67	72	98	68	82
Llama 4 Maverick	Meta 🇺🇸	Open	75.5	80	52	98	72	78
Muse Glimmer 30B	Meta 🇺🇸	Open	75.5	66	72	95	70	82
Gemma 4 31B Dense	Google 🇺🇸	Open	75.0	83	74	91	64	60
Devstral Small 2505	Mistral AI 🇫🇷	Open	74.9	69	80	92	66	72
Claude Sonnet 4.6	Anthropic 🇺🇸	Closed	74.6	84	78	97	62	50
Qwen2.5-VL-72B-Instruct	Alibaba 🇨🇳	Open	74.4	74	58	98	72	75
Claude Opus 4	Anthropic 🇺🇸	Closed	74.2	83	78	76	66	62
Gemini 2.5 Flash	Google 🇺🇸	Closed	73.9	80	67	99	68	57
Gemma 4 26B MoE	Google 🇺🇸	Open	73.9	81	73	94	62	58
GPT-realtime	OpenAI 🇺🇸	Closed	73.9	78	67	99	72	57
Mistral Small 4	Mistral AI 🇫🇷	Open	73.4	80	74	92	62	57
Grok 3	SpaceXAI 🇺🇸	Closed	73.3	84	62	79	75	63
GPT-5 mini	OpenAI 🇺🇸	Closed	73.0	78	70	100	68	52
GPT-5.2 Instant	OpenAI 🇺🇸	Closed	73.0	80	65	100	69	53
OpenAI o3-mini	OpenAI 🇺🇸	Open	72.9	81	61	99	68	56
Claude Sonnet 4.5	Anthropic 🇺🇸	Closed	72.5	82	77	92	60	48
Gemini 3 Deep Think	Google 🇺🇸	Closed	72.4	94	74	63	59	55
Llama 4 Scout	Meta 🇺🇸	Open	72.3	76	46	98	69	76
NVIDIA Nemotron 3 Nano 30B A3B	NVIDIA 🇺🇸	Open	72.3	66	70	97	61	74
Gemini 3.5 Flash Cyber	Google 🇺🇸	Closed	72.2	78	84	25	92	72
GPT-5.1 Instant	OpenAI 🇺🇸	Closed	71.6	79	64	100	67	50
OpenAI o1 2024-12-05	OpenAI 🇺🇸	Closed	71.6	82	55	94	68	58
Claude Haiku 4.5	Anthropic 🇺🇸	Closed	71.3	75	76	100	60	48
Gemini 3.5 Flash-Lite	Google 🇺🇸	Open	71.1	72	76	99	60	52
Gemini 2.5 Flash Preview 04-17	Google 🇺🇸	Closed	70.9	78	62	89	68	58
Qwen3-32B	Alibaba 🇨🇳	Open	70.8	72	48	98	67	74
Qwen3-30B-A3B	Alibaba 🇨🇳	Open	70.6	70	52	98	66	73
OpenAI o1 2024-12-17	OpenAI 🇺🇸	Closed	70.5	83	58	78	70	59
Gemini 2.0 Pro Experimental	Google 🇺🇸	Closed	70.4	83	61	74	69	59
Command A+	Cohere 🇨🇦	Open	70.2	73	62	84	70	64
Gemini 2.0 Flash	Google 🇺🇸	Open	70.2	78	55	99	65	56
Claude 3.5 Sonnet 2024-10-22	Anthropic 🇺🇸	Closed	70.1	77	60	98	63	53
Mistral Small 3.2	Mistral AI 🇫🇷	Open	70.1	70	54	96	63	72
Grok 3 mini	SpaceXAI 🇺🇸	Closed	69.7	77	64	78	70	58
Magistral Medium 2506	Mistral AI 🇫🇷	Open	69.7	78	63	88	66	53
Command A 03-2025	Cohere 🇨🇦	Open	69.4	75	62	92	61	57
Isaac GR00T N1	NVIDIA 🇺🇸	Open	69.4	62	72	92	55	71
GPT-4.1 mini	OpenAI 🇺🇸	Closed	68.9	72	58	99	64	56
DeepSeek-V3	DeepSeek 🇨🇳	Open	67.9	75	24	94	73	77
Gemini Robotics ER 2	Google 🇺🇸	Closed	67.9	72	84	48	66	62
Llama 3.1 405B Instruct	Meta 🇺🇸	Open	67.8	80	24	82	75	77
Mistral Medium 3	Mistral AI 🇫🇷	Closed	67.8	75	58	96	62	50
Llama 3.3 70B Instruct	Meta 🇺🇸	Open	67.7	75	22	98	72	76
Magistral Small 2506	Mistral AI 🇫🇷	Open	67.6	66	50	98	62	69
DeepSeek-R1	DeepSeek 🇨🇳	Open	67.4	74	45	88	66	65
Gemini 2.0 Flash Experimental	Google 🇺🇸	Closed	67.4	77	52	88	64	55
Llama 3.2 90B Vision Instruct	Meta 🇺🇸	Open	67.3	74	32	88	70	75
Kimi K2 Base	Moonshot AI 🇨🇳	Open	67.2	78	25	91	70	73
GPT-4o Realtime Preview 2024-10-01	OpenAI 🇺🇸	Closed	67.1	74	53	88	67	55
GPT-4.5 Preview	OpenAI 🇺🇸	Closed	67.0	82	44	78	69	57
ROME	Alibaba 🇨🇳	Open	66.9	58	72	68	64	76
Claude Sonnet 4	Anthropic 🇺🇸	Closed	66.7	78	65	82	56	48
Qwen2.5-Max	Alibaba 🇨🇳	Open	66.4	80	35	90	69	57
Grok-2 Vision 1212	SpaceXAI 🇺🇸	Closed	66.3	76	31	94	73	61
GPT-Live-1	OpenAI 🇺🇸	Closed	65.9	76	72	87	48	42
Gemini Robotics 1.5	Google 🇺🇸	Closed	65.6	72	82	38	60	64
Gemini Robotics On-Device	Google 🇺🇸	Closed	65.4	68	80	45	58	67
GPT-4o 2024-08-06	OpenAI 🇺🇸	Open	65.2	75	44	97	62	50
o1-preview 2024-09-12	OpenAI 🇺🇸	Closed	65.0	78	51	72	65	54
Grok-2 1212	SpaceXAI 🇺🇸	Open	64.7	74	28	96	71	59
Qwen2-VL-72B-Instruct	Alibaba 🇨🇳	Closed	64.7	70	38	85	65	68
Qwen3-14B	Alibaba 🇨🇳	Open	64.5	64	38	98	60	70
GPT-Image-2.5 Sunburst	OpenAI 🇺🇸	Closed	64.2	86	4	90	82	59
Qwen2.5-72B-Instruct	Alibaba 🇨🇳	Open	64.2	72	18	94	68	73
Alpamayo 1	NVIDIA 🇺🇸	Open	64.1	60	76	58	55	69
Llama 3.1 70B Instruct	Meta 🇺🇸	Open	64.1	71	18	96	68	73
Veo 3	Google 🇺🇸	Closed	64.1	82	8	84	83	64
Gemini Robotics 2	Google 🇺🇸	Closed	63.6	70	82	24	64	65
QVQ-72B-Preview	Alibaba 🇨🇳	Open	63.6	74	24	78	69	72
Claude 3.5 Sonnet	Anthropic 🇺🇸	Closed	63.5	75	40	96	59	48
GPT-4o	OpenAI 🇺🇸	Closed	63.1	73	39	96	60	49
GPT-Image-2.5 Flare	OpenAI 🇺🇸	Closed	63.0	82	4	96	79	56
Mistral Small 3.1 2503	Mistral AI 🇫🇷	Open	63.0	65	32	98	58	68
Gemini Robotics	Google 🇺🇸	Closed	62.7	75	78	25	56	62
Qwen2.5-Omni-7B	Alibaba 🇨🇳	Open	62.7	63	36	98	57	66
Gemma 4 E4B	Google 🇺🇸	Open	62.6	67	55	98	45	50
GPT-5.4 nano	OpenAI 🇺🇸	Closed	62.5	72	58	94	45	42
Gemma 3 27B IT	Google 🇺🇸	Open	62.0	66	22	98	62	69
Mistral Large 2 24.07	Mistral AI 🇫🇷	Open	61.7	73	20	82	65	68
Qwen-Image	Alibaba 🇨🇳	Open	61.7	68	3	99	76	72
DeepSeek-Coder-V2-Instruct	DeepSeek 🇨🇳	Open	61.3	71	20	78	70	69
Gemini 1.5 Pro-002	Google 🇺🇸	Closed	61.2	76	24	92	63	52
Gemini 2.5 Flash-Lite Preview 06-17	Google 🇺🇸	Closed	61.2	66	45	96	56	48
Midjourney V1 Video Model	Midjourney 🇺🇸	Closed	60.8	74	3	94	80	59
o1-mini 2024-09-12	OpenAI 🇺🇸	Closed	60.7	69	45	88	57	46
Gemini Robotics-ER	Google 🇺🇸	Closed	60.6	72	70	35	54	58
Qwen2 72B Instruct	Alibaba 🇨🇳	Open	60.6	69	14	89	66	69
Llama 3 70B Instruct	Meta 🇺🇸	Open	60.5	67	14	94	63	70
Imagen 4	Google 🇺🇸	Closed	60.4	78	4	88	76	58
Gemini Robotics On-Device 2	Google 🇺🇸	Closed	60.2	62	76	25	60	68
Qwen3-8B	Alibaba 🇨🇳	Open	60.2	58	32	98	55	67
Grok-2	SpaceXAI 🇺🇸	Closed	60.0	72	24	79	68	57
Amazon Nova Pro	Amazon 🇺🇸	Open	59.7	70	24	91	63	54
Llama 3.1 405B Base	Meta 🇺🇸	Open	59.2	75	6	74	67	72
Llama 3.2 11B Vision Instruct	Meta 🇺🇸	Open	59.1	59	24	98	58	66
Claude 3 Opus	Anthropic 🇺🇸	Closed	58.6	72	38	76	56	48
Gemini 2.0 Flash-Lite	Google 🇺🇸	Closed	58.4	65	34	95	56	47
Gemma 2 27B IT	Google 🇺🇸	Open	58.4	64	12	95	61	67
Llama 3.1 70B Base	Meta 🇺🇸	Open	57.7	66	5	94	60	69
Llama 3.2 90B Vision Base	Meta 🇺🇸	Open	57.7	68	6	83	64	70
Qwen2-VL-7B-Instruct	Alibaba 🇨🇳	Open	57.3	55	25	98	55	64
Gemma 4 E2B	Google 🇺🇸	Open	57.2	60	48	98	38	46
Gemma 3 12B IT	Google 🇺🇸	Open	57.1	59	17	99	54	65
GPT-4 Turbo 1106 Preview	OpenAI 🇺🇸	Closed	56.7	68	38	78	52	46
GPT-4.1 nano	OpenAI 🇺🇸	Closed	56.6	58	38	99	49	47
gpt-4-0125-preview	OpenAI 🇺🇸	Closed	56.2	68	34	82	51	45
Qwen2 57B-A14B Instruct	Alibaba 🇨🇳	Open	56.0	62	11	90	59	64
Amazon Nova Sonic	Amazon 🇺🇸	Closed	55.8	58	32	95	54	48
Gemma 3n	Google 🇺🇸	Open	55.8	58	12	96	53	68
Llama 3 70B Base	Meta 🇺🇸	Open	55.6	63	5	93	56	67
DeepSeek-Coder-V2-Lite-Instruct	DeepSeek 🇨🇳	Open	55.4	55	15	93	61	64
Llama 2 70B Chat	Meta 🇺🇸	Open	55.4	59	17	88	55	64
Qwen2 72B Base	Alibaba 🇨🇳	Open	54.8	64	5	86	58	65
Llama 3.1 8B Instruct	Meta 🇺🇸	Open	54.6	54	12	98	54	66
Command R7B 12-2024	Cohere 🇨🇦	Open	54.3	54	25	94	50	57
Mixtral 8x7B Instruct v0.1	Mistral AI 🇫🇷	Open	54.3	61	7	87	65	59
Qwen-72B-Chat	Alibaba 🇨🇳	Open	54.3	60	10	88	55	64
Gemini 1.5 Flash-002	Google 🇺🇸	Closed	54.2	62	19	98	54	45
Gemma 3 27B Base	Google 🇺🇸	Open	54.2	58	6	97	55	64
GPT-4-0613	OpenAI 🇺🇸	Closed	54.2	65	38	72	50	44
Qwen3-4B	Alibaba 🇨🇳	Open	54.2	50	24	98	47	63
Qwen2-Audio-7B-Instruct	Alibaba 🇨🇳	Open	54.0	55	12	97	54	62
Grok-2 mini	SpaceXAI 🇺🇸	Closed	53.9	62	18	82	60	52
Amazon Nova Lite	Amazon 🇺🇸	Closed	53.6	60	18	94	54	49
DeepSeek-Coder-V2-Base	DeepSeek 🇨🇳	Open	53.6	66	5	72	61	64
Claude 3 Sonnet	Anthropic 🇺🇸	Closed	53.3	63	30	90	47	39
Mistral NeMo 12B Instruct	Mistral AI 🇫🇷	Open	53.1	53	12	97	51	63
Mixtral 8x7B v0.1	Mistral AI 🇫🇷	Open	52.9	58	5	88	62	60
Gemini 1.5 Pro	Google 🇺🇸	Closed	52.8	74	18	48	62	52
Gemma 2 27B Base	Google 🇺🇸	Open	52.7	58	4	94	52	63
Code Llama 70B Instruct	Meta 🇺🇸	Open	52.5	55	8	94	54	61
GPT-Red	OpenAI 🇺🇸	Internal	52.4	66	50	5	92	40
Gemma 2 9B IT	Google 🇺🇸	Open	52.3	53	9	97	52	61
Mistral Large 1.0	Mistral AI 🇫🇷	Open	52.3	66	12	80	58	47
Phi-4	Microsoft 🇺🇸	Open	52.2	63	16	70	55	56
GPT-4-32k-0613	OpenAI 🇺🇸	Closed	51.8	66	39	48	52	46
Gemma 3 4B IT	Google 🇺🇸	Open	51.7	49	14	100	45	62
Aurora	SpaceXAI 🇺🇸	Closed	51.6	60	6	82	68	49
Llama 3 8B Instruct	Meta 🇺🇸	Open	51.5	50	10	97	50	62
GPT-4	OpenAI 🇺🇸	Closed	51.4	64	32	68	48	42
Llama 2 70B Base	Meta 🇺🇸	Open	51.4	55	6	86	52	65
Code Llama 34B	Meta 🇺🇸	Open	51.2	52	10	92	52	60
Nemotron-4 340B Instruct	NVIDIA 🇺🇸	Open	51.2	68	12	58	63	51
Gemini 1.5 Flash	Google 🇺🇸	Open	51.0	58	16	94	51	43
Qwen2 57B-A14B Base	Alibaba 🇨🇳	Open	50.8	57	4	88	51	60
Llama 3.2 11B Vision Base	Meta 🇺🇸	Open	50.7	52	5	96	50	61
Qwen-72B	Alibaba 🇨🇳	Open	50.7	57	4	87	50	61
Claude 3.5 Haiku	Anthropic 🇺🇸	Closed	50.6	58	34	88	36	39
Gemma 3 12B Base	Google 🇺🇸	Open	50.6	52	5	98	47	61
FLUX.1 [dev]	Black Forest Labs 🇩🇪	Open	50.4	51	1	92	61	60
Phi-3 Mini 128K Instruct	Microsoft 🇺🇸	Open	50.4	49	13	96	47	58
FLUX.1 [schnell]	Black Forest Labs 🇩🇪	Open	50.3	45	1	99	58	65
Sora Turbo	OpenAI 🇺🇸	Closed	50.0	64	7	72	66	43
Gemini 1.0 Ultra	Google 🇺🇸	Closed	49.9	72	8	54	60	48
Code Llama 70B Python	Meta 🇺🇸	Open	49.8	53	4	94	48	59
Qwen2 7B Instruct	Alibaba 🇨🇳	Open	49.8	49	8	96	50	58
Llama 2 13B Chat	Meta 🇺🇸	Open	49.6	48	12	95	46	58
Llama 3.2 3B Instruct	Meta 🇺🇸	Open	49.6	44	15	100	44	59
Llama 3.1 8B Base	Meta 🇺🇸	Open	49.5	49	4	97	47	62
Gemini 1.0 Pro	Google 🇺🇸	Closed	49.4	64	6	81	54	43
GPT-3.5 Turbo-0613	OpenAI 🇺🇸	Closed	49.0	49	31	90	48	37
Qwen3-1.7B	Alibaba 🇨🇳	Open	48.9	42	18	98	40	60
DeepSeek-Coder-V2-Lite-Base	DeepSeek 🇨🇳	Open	48.8	49	4	91	52	59
Mistral 7B 0.1 Instruct	Mistral AI 🇫🇷	Open	48.8	47	8	97	42	61
Code Llama 70B	Meta 🇺🇸	Open	48.6	51	4	94	45	58
GPT-4o mini 2024-07-18	OpenAI 🇺🇸	Open	48.5	54	14	99	47	38
Qwen2-Audio-7B	Alibaba 🇨🇳	Open	48.5	50	3	95	47	58
Gemma 7B IT	Google 🇺🇸	Open	48.4	47	7	95	48	57
Mistral NeMo 12B Base	Mistral AI 🇫🇷	Open	48.2	48	3	96	45	60
Qwen2-VL-2B-Instruct	Alibaba 🇨🇳	Open	48.1	40	15	98	45	58
Llama 3 8B Base	Meta 🇺🇸	Open	47.6	46	4	96	44	60
Phi-3 Mini 4K Instruct	Microsoft 🇺🇸	Open	47.6	47	8	96	43	55
Amazon Nova Micro	Amazon 🇺🇸	Open	47.5	50	10	96	46	46
Claude 2.1	Anthropic 🇺🇸	Closed	47.4	58	24	75	42	38
FLUX.1 [pro]	Black Forest Labs 🇩🇪	Open	47.2	55	1	78	60	49
Gemma 2 9B Base	Google 🇺🇸	Open	47.1	47	3	96	44	57
Llama 2 13B Base	Meta 🇺🇸	Open	46.8	45	4	94	44	59
Waymo Driver 5th generation	Waymo 🇺🇸	Internal	46.4	47	84	38	8	43
Qwen-7B-Chat	Alibaba 🇨🇳	Open	46.1	45	10	90	43	53
Llama 2 7B Chat	Meta 🇺🇸	Open	45.8	42	9	96	41	54
Claude 2	Anthropic 🇺🇸	Closed	45.7	58	19	70	43	37
Gemma 3 4B Base	Google 🇺🇸	Open	45.7	42	4	99	39	58
Mistral 7B 0.1 Base	Mistral AI 🇫🇷	Open	45.7	44	4	96	38	58
PaLM 2	Google 🇺🇸	Closed	45.7	56	8	86	43	39
Veo 2	Google 🇺🇸	Closed	45.3	69	6	30	65	45
gpt-oss-safeguard-120b	OpenAI 🇺🇸	Open	44.8	45	20	99	32	38
Qwen2 7B Base	Alibaba 🇨🇳	Open	44.8	43	3	95	42	54
Waymo Driver 4th generation	Waymo 🇺🇸	Closed	44.5	43	80	44	8	39
Claude 3 Haiku	Anthropic 🇺🇸	Closed	44.4	49	20	96	36	29
Gemma 7B Base	Google 🇺🇸	Open	44.4	42	3	95	41	54
Gemma 3 1B IT	Google 🇺🇸	Open	44.3	38	10	100	34	54
Mistral Small 1.0	Mistral AI 🇫🇷	Closed	44.3	50	7	88	46	39
Nemotron-4 340B Base	NVIDIA 🇺🇸	Open	44.1	64	4	43	55	47
Mistral OCR 2503	Mistral AI 🇫🇷	Closed	43.5	52	15	84	35	35
Llama 2 7B Base	Meta 🇺🇸	Open	43.4	39	3	95	39	55
GPT-3	OpenAI 🇺🇸	Closed	43.1	55	8	68	48	38
Qwen-7B	Alibaba 🇨🇳	Open	42.7	42	4	88	39	51
Llama 3.2 3B Base	Meta 🇺🇸	Open	42.3	36	3	99	35	54
Gemma 2 2B IT	Google 🇺🇸	Open	41.9	35	6	99	34	51
Llama 3.2 1B Instruct	Meta 🇺🇸	Open	41.6	31	9	100	34	52
Gemma 2B IT	Google 🇺🇸	Open	41.4	34	5	96	37	51
StableLM-Base-Alpha-7B	Stability AI 🇬🇧	Open	40.7	42	4	78	40	48
Qwen2 1.5B Instruct	Alibaba 🇨🇳	Open	39.9	32	5	97	34	48
BLOOM 176B	Hugging Face 🇺🇸	Open	39.8	47	2	68	39	46
DALL-E 3	OpenAI 🇺🇸	Closed	39.8	47	3	68	55	33
Gemma 3 1B Base	Google 🇺🇸	Open	39.4	32	2	99	30	50
gpt-oss-safeguard-20b	OpenAI 🇺🇸	Open	39.4	35	16	99	27	34
ERNIE 4.0	Baidu 🇨🇳	Closed	39.2	60	11	34	43	37
North Small Translate 1.0	Cohere 🇨🇦	Open	39.0	50	25	51	25	38
GPT-NeoX-20B	EleutherAI 🇺🇸	Open	38.9	42	5	76	31	46
Claude Instant 1.2	Anthropic 🇺🇸	Closed	38.8	45	13	72	35	33
Gemma 2B Base	Google 🇺🇸	Open	38.1	30	2	96	31	48
Dolly 2.0 12B	Databricks 🇺🇸	Open	37.9	42	4	68	39	42
Gemma 2 2B Base	Google 🇺🇸	Open	37.6	29	2	98	28	48
Defense Llama	Scale AI 🇺🇸	Open	37.5	44	15	23	72	35
StableLM-Base-Alpha-3B	Stability AI 🇬🇧	Open	37.4	37	3	78	36	43
Jurassic-1 Jumbo	AI21 Labs 🇮🇱	Closed	36.4	47	3	58	40	35
Llama 3.2 1B Base	Meta 🇺🇸	Open	36.3	25	2	99	26	48
Qwen2 1.5B Base	Alibaba 🇨🇳	Open	36.1	27	2	96	28	45
GPT-3 175B	OpenAI 🇺🇸	Closed	35.9	48	10	45	42	32
Inflection-1	Inflection AI 🇺🇸	Closed	35.4	48	5	58	34	31
Qwen2 0.5B Instruct	Alibaba 🇨🇳	Open	35.4	24	3	97	27	45
DALL-E 2	OpenAI 🇺🇸	Closed	35.1	37	1	72	46	30
ERNIE 3.0 Titan	Baidu 🇨🇳	Closed	35.1	52	8	36	39	33
Grok-1	SpaceXAI 🇺🇸	Open	34.0	52	7	18	48	35
BlenderBot 3 175B	Meta 🇺🇸	Closed	33.8	45	24	38	24	31
code-davinci-002	OpenAI 🇺🇸	Closed	33.8	46	5	57	35	26
NVIDIA Megatron 530B	NVIDIA 🇺🇸	Closed	33.5	56	4	24	42	30
WebGPT 175B	OpenAI 🇺🇸	Internal	33.0	52	32	3	38	24
Qwen2 0.5B Base	Alibaba 🇨🇳	Open	32.3	20	1	96	22	42
BlenderBot 2.0 2.7B	Meta 🇺🇸	Open	32.1	38	22	42	28	29
PanGu-alpha 200B	Huawei 🇨🇳	Open	31.7	50	8	28	34	29
Nemotron-4 340B Reward	NVIDIA 🇺🇸	Open	30.9	46	2	32	38	31
LLaMA 7B	Meta 🇺🇸	Open	30.5	45	4	25	39	33
PaLM 540B	Google 🇺🇸	Internal	28.6	55	4	4	35	27
Chinchilla 70B	Google 🇺🇸	Internal	27.6	55	5	8	30	22
BlenderBot 2.0 400M	Meta 🇺🇸	Open	27.5	30	18	45	22	24
OPT-175B	Meta 🇺🇸	Open	27.4	45	4	25	30	24
GPT-2	OpenAI 🇺🇸	Open	26.1	22	2	68	27	24
Turing-NLG 17B	Microsoft 🇺🇸	Closed	25.5	39	3	58	14	12
Codex-12B	OpenAI 🇺🇸	Internal	25.2	45	4	18	28	20
Qwen3Guard-Gen-8B	Alibaba 🇨🇳	Open	22.3	23	1	85	8	6
LaMDA 137B	Google 🇺🇸	Internal	21.4	42	2	3	24	22
Qwen3Guard-Stream-8B	Alibaba 🇨🇳	Open	21.2	20	2	85	6	6
BlenderBot 9.4B	Meta 🇺🇸	Open	20.8	34	2	30	18	15
Qwen3Guard-Gen-4B	Alibaba 🇨🇳	Open	20.3	18	1	85	6	5
Shieldstral 1.0 3B	Mistral AI 🇫🇷	Open	20.1	20	1	78	8	5
Qwen3Guard-Stream-4B	Alibaba 🇨🇳	Open	19.3	15	2	85	4	5
Qwen3Guard-Gen-0.6B	Alibaba 🇨🇳	Open	16.8	10	1	82	3	4
Qwen3Guard-Stream-0.6B	Alibaba 🇨🇳	Open	16.2	8	2	82	2	4

FREQUENTLY ASKED QUESTIONS

Questions about this page

Short answers to common questions about the page, its evidence, and the limits of what DoomBench claims.

What is a Model Doom Score?

It is a version-specific profile across capability, autonomy, deployment, misuse potential, and control difficulty. It is not the overall Doom Index and it is not a probability of catastrophe.

Does a model score directly move the overall index?

No. Only accepted, dated evidence items move the overall index. A model release or evaluation matters there through its separately reviewed evidence assessment.

What do Open, Closed, Internal, and Unknown mean?

Open identifies publicly available artifacts such as weights or source materials, Closed identifies proprietary hosted access, Internal identifies systems not publicly released, and Unknown means the current sources do not support a confident classification.

Why can a model score be higher than the overall index?

The values answer different questions. A model score rates one technical profile, while the overall index aggregates the complete, time-weighted real-world evidence record.

Questions about this page

Share this page