Google logo

COMPANY PROFILE

Google 🇺🇸

DoomBench currently associates 68 models and 128 evidence assessments with Google. Company attribution is separate from each model's version-specific risk score.

Tracked models: 68
Evidence items: 128
Toward-doom share: 12.2%
Net index pressure: +7.30

0 comments · 0 votesOpen discussion

No comments yet. Start the discussion.

VERSION-SPECIFIC PROFILES

Models from Google

Availability distinguishes public artifacts, proprietary hosted access, internal systems, and profiles whose current source evidence remains inconclusive.

Filter modelsAvailabilityRelease yearRelease monthOrder68 models · average 61.2

92.8ClosedGemini 4 Argon 86.7ClosedGemini 3.8 Flash 84.4UnknownGemini 4 Carbon 82.9ClosedGemini 3.7 Flash 81.5ClosedGemini 3 Pro 81.4OpenGemini 2.5 Pro 80.9ClosedGemini 3.5 Flash 80.7ClosedGemini 3.1 Pro 79.7ClosedGemini 2.5 Computer Use 79.7ClosedGemini 3.8 Flash Cyber 79.1ClosedGemini 2.5 Pro Preview 06-05 78.7OpenGemini 3.6 Flash 78.3ClosedGemini 3 Flash 77.5ClosedGemini 2.5 Pro Preview 05-06 76.2ClosedGemini 2.5 Pro Experimental 75.0OpenGemma 4 31B Dense 73.9ClosedGemini 2.5 Flash 73.9OpenGemma 4 26B MoE 72.3ClosedGemini 3 Deep Think 72.2ClosedGemini 3.5 Flash Cyber 71.3ClosedGemini Robotics ER 2 71.1OpenGemini 3.5 Flash-Lite 70.9ClosedGemini 2.5 Flash Preview 04-17 70.4ClosedGemini 2.0 Pro Experimental 70.2OpenGemini 2.0 Flash 67.4ClosedGemini 2.0 Flash Experimental 65.6ClosedGemini Robotics 1.5 65.4ClosedGemini Robotics On-Device 64.1ClosedVeo 3 63.6ClosedGemini Robotics 2 62.7ClosedGemini Robotics 62.6OpenGemma 4 E4B 62.0OpenGemma 3 27B IT 61.2ClosedGemini 1.5 Pro-002 61.2ClosedGemini 2.5 Flash-Lite Preview 06-17 60.6ClosedGemini Robotics-ER 60.4ClosedImagen 4 60.2ClosedGemini Robotics On-Device 2 58.4ClosedGemini 2.0 Flash-Lite 58.4OpenGemma 2 27B IT 57.2OpenGemma 4 E2B 57.1OpenGemma 3 12B IT 55.7OpenGemma 3n 54.2ClosedGemini 1.5 Flash-002 54.2OpenGemma 3 27B Base 52.8ClosedGemini 1.5 Pro 52.7OpenGemma 2 27B Base 52.3OpenGemma 2 9B IT 51.7OpenGemma 3 4B IT 51.0OpenGemini 1.5 Flash 50.6OpenGemma 3 12B Base 49.9ClosedGemini 1.0 Ultra 49.4ClosedGemini 1.0 Pro 48.4OpenGemma 7B IT 47.1OpenGemma 2 9B Base 45.7OpenGemma 3 4B Base 45.6ClosedPaLM 2 45.3ClosedVeo 2 44.4OpenGemma 7B Base 44.3OpenGemma 3 1B IT 41.9OpenGemma 2 2B IT 41.4OpenGemma 2B IT 39.4OpenGemma 3 1B Base 38.1OpenGemma 2B Base 37.6OpenGemma 2 2B Base 28.6InternalPaLM 540B 27.6InternalChinchilla 70B 21.4InternalLaMDA 137B

Filtered company model comparison
Model	Availability	Released	Doom Score
Gemini 4 Argon	Closed	30 Sept 2026	92.8
Gemini 3.8 Flash	Closed	02 Sept 2026	86.7
Gemini 4 Carbon	Unknown	28 Oct 2026	84.4
Gemini 3.7 Flash	Closed	13 Aug 2026	82.9
Gemini 3 Pro	Closed	18 Nov 2025	81.5
Gemini 2.5 Pro	Open	17 Jun 2025	81.4
Gemini 3.5 Flash	Closed	19 May 2026	80.9
Gemini 3.1 Pro	Closed	19 Feb 2026	80.7
Gemini 2.5 Computer Use	Closed	07 Oct 2025	79.7
Gemini 3.8 Flash Cyber	Closed	02 Sept 2026	79.7
Gemini 2.5 Pro Preview 06-05	Closed	05 Jun 2025	79.1
Gemini 3.6 Flash	Open	21 Jul 2026	78.7
Gemini 3 Flash	Closed	17 Dec 2025	78.3
Gemini 2.5 Pro Preview 05-06	Closed	06 May 2025	77.5
Gemini 2.5 Pro Experimental	Closed	25 Mar 2025	76.2
Gemma 4 31B Dense	Open	02 Apr 2026	75.0
Gemini 2.5 Flash	Closed	17 Jun 2025	73.9
Gemma 4 26B MoE	Open	02 Apr 2026	73.9
Gemini 3 Deep Think	Closed	12 Feb 2026	72.3
Gemini 3.5 Flash Cyber	Closed	21 Jul 2026	72.2
Gemini Robotics ER 2	Closed	30 Jul 2026	71.3
Gemini 3.5 Flash-Lite	Open	21 Jul 2026	71.1
Gemini 2.5 Flash Preview 04-17	Closed	17 Apr 2025	70.9
Gemini 2.0 Pro Experimental	Closed	05 Feb 2025	70.4
Gemini 2.0 Flash	Open	05 Feb 2025	70.2
Gemini 2.0 Flash Experimental	Closed	11 Dec 2024	67.4
Gemini Robotics 1.5	Closed	25 Sept 2025	65.6
Gemini Robotics On-Device	Closed	24 Jun 2025	65.4
Veo 3	Closed	20 May 2025	64.1
Gemini Robotics 2	Closed	30 Jul 2026	63.6
Gemini Robotics	Closed	12 Mar 2025	62.7
Gemma 4 E4B	Open	02 Apr 2026	62.6
Gemma 3 27B IT	Open	12 Mar 2025	62.0
Gemini 1.5 Pro-002	Closed	24 Sept 2024	61.2
Gemini 2.5 Flash-Lite Preview 06-17	Closed	17 Jun 2025	61.2
Gemini Robotics-ER	Closed	12 Mar 2025	60.6
Imagen 4	Closed	20 May 2025	60.4
Gemini Robotics On-Device 2	Closed	30 Jul 2026	60.2
Gemini 2.0 Flash-Lite	Closed	05 Feb 2025	58.4
Gemma 2 27B IT	Open	27 Jun 2024	58.4
Gemma 4 E2B	Open	02 Apr 2026	57.2
Gemma 3 12B IT	Open	12 Mar 2025	57.1
Gemma 3n	Open	20 May 2025	55.7
Gemini 1.5 Flash-002	Closed	24 Sept 2024	54.2
Gemma 3 27B Base	Open	12 Mar 2025	54.2
Gemini 1.5 Pro	Closed	15 Feb 2024	52.8
Gemma 2 27B Base	Open	27 Jun 2024	52.7
Gemma 2 9B IT	Open	27 Jun 2024	52.3
Gemma 3 4B IT	Open	12 Mar 2025	51.7
Gemini 1.5 Flash	Open	14 May 2024	51.0
Gemma 3 12B Base	Open	12 Mar 2025	50.6
Gemini 1.0 Ultra	Closed	08 Feb 2024	49.9
Gemini 1.0 Pro	Closed	06 Dec 2023	49.4
Gemma 7B IT	Open	21 Feb 2024	48.4
Gemma 2 9B Base	Open	27 Jun 2024	47.1
Gemma 3 4B Base	Open	12 Mar 2025	45.7
PaLM 2	Closed	10 May 2023	45.6
Veo 2	Closed	16 Dec 2024	45.3
Gemma 7B Base	Open	21 Feb 2024	44.4
Gemma 3 1B IT	Open	12 Mar 2025	44.3
Gemma 2 2B IT	Open	31 Jul 2024	41.9
Gemma 2B IT	Open	21 Feb 2024	41.4
Gemma 3 1B Base	Open	12 Mar 2025	39.4
Gemma 2B Base	Open	21 Feb 2024	38.1
Gemma 2 2B Base	Open	31 Jul 2024	37.6
PaLM 540B	Internal	05 Apr 2022	28.6
Chinchilla 70B	Internal	29 Mar 2022	27.6
LaMDA 137B	Internal	20 Jan 2022	21.4

RELATED EVIDENCE

Latest assessments involving Google

View all 128 items →

28 Oct 2026Capability TOWARD 40

Google announces Gemini 4 Carbon for restricted cyber-defender rollout

Google announced Gemini 4 Carbon with stronger cyber benchmarks and reported multi-day engineering workflows. Initial access is planned for selected cyber defenders; broader access remains pending. Its claimed reductions in misalignment and prompt-injection success are bounded evaluations, not proof of general safety.

Google 🇺🇸Gemini 4 Carbon

Google Blog

28 Oct 2026Governance AWAY 18

House committee leaders ask AI companies for rapid reporting of misalignment incidents

House Democratic committee leaders asked five frontier-AI companies for past incident information and future reporting within 48 hours, or 24 hours for specified critical-infrastructure incidents. These were congressional requests, not a newly enacted mandatory reporting law.

Anthropic 🇺🇸Google 🇺🇸Meta 🇺🇸OpenAI 🇺🇸SpaceXAI 🇺🇸

House Energy and Commerce Democrats

28 Oct 2026Labor TOWARD 20

Nevada labor reporting shows AI substitution alongside data-center job gains

Reuters reports about 200 MGM call-center layoffs in 2025, attributed to a Nevada union official, alongside growing construction work at AI data centers. It also cites 80% of MGM guest interactions moving through AI or digital platforms. These mixed observations do not establish net job losses; projected hospitality losses by 2035 remain forecasts.

Google 🇺🇸MGM Resorts International 🌐Switch 🌐Tract 🌐Waymo 🇺🇸Zoox 🇺🇸

Reuters

27 Oct 2026Autonomy TOWARD 24

Google opens Gemini Robotics ER 2 public preview on its cloud platform

Gemini Robotics ER 2 became available in public preview through Google Cloud, extending access to its spatial reasoning, task verification and high-level orchestration of robot-control tools. The announcement does not establish unrestricted safe deployment or a new underlying model version.

Google 🇺🇸Gemini Robotics ER 2

Google Cloud

27 Oct 2026Capability TOWARD 36

Anthropic reports five prompted exploit chains across browsers and operating systems

Anthropic described an unnamed internal Claude research model producing exploit chains after human prompting for Chrome, Safari, Android, Windows and Linux. It reported coordinated fixes; Google's October 20 Chrome advisory independently credits Claude-assisted findings. These are controlled demonstrations, not a new autonomous escape.

Anthropic 🇺🇸Apple 🇺🇸Google 🇺🇸Microsoft 🇺🇸

Anthropic Research

21 Oct 2026Deployment TOWARD 18

Google Cloud starts private preview of OpenAI frontier models on its enterprise platform

Google Cloud began a select-customer private preview of GPT-6 Astra, GPT-6.1 Sol and GPT-6 Luna in Model Garden. Customers can access them through Google Cloud APIs and connected workflows. Zero Data Retention and Confidential Compute were described as coming later.

Google 🇺🇸OpenAI 🇺🇸GPT-6 Astra GPT-6 Luna GPT-6.1 Sol

Google Cloud

13 Oct 2026Deployment TOWARD 35

Google expands Gemini 4 Argon from restricted access to paid APIs and consumer plans

Google announced a broader Gemini 4 Argon rollout through the paid Gemini API, Ultra access in Antigravity and AI Studio, and Pro or Ultra consumer plans. The frontier model was previously limited to internal use and trusted defenders. Cyber access without guardrails remains separately vetted through Fairwind; the rollout announcement does not prove availability for every account.

Google 🇺🇸Gemini 4 Argon

Google

08 Oct 2026Race TOWARD 24

Bengio urges safety-focused researchers to leave frontier labs for independent work

Yoshua Bengio argues that commercial frontier labs subordinate safety to competitive deployment incentives and urges safety-focused researchers to move to independent institutions. This is his dated argument and proposed response, not evidence that a mass departure occurred.

Anthropic 🇺🇸Google 🇺🇸OpenAI 🇺🇸Yoshua Bengio 🇫🇷

Yoshua Bengio / Transformer

08 Oct 2026Autonomy TOWARD 30

Google introduces persistent Gemini coworker agents with enterprise control boundaries

Google announced a universal enterprise agent that can run objectives for hours or days, create sub-agents and operate through dedicated coworker identities. The platform combines persistent memory and tool access with per-agent authorization, audit trails, sandboxing and a network-policy gateway.

Anthropic 🇺🇸Google 🇺🇸

Google Cloud

07 Oct 2026Deployment TOWARD 55

Anthropic releases Haiku 5.5 with cheaper computer-use and coding agents

Anthropic released Claude Haiku 5.5 across its API, Amazon Web Services, Google Cloud and Microsoft Azure. It reports stronger computer-use and coding results than Haiku 4.5 and about 75% lower average task cost, after accounting for token usage. Broader inexpensive agent access is established; benchmark gains are provider-reported, not proof of job displacement.

Amazon 🇺🇸Anthropic 🇺🇸Google 🇺🇸Microsoft 🇺🇸Claude Haiku 5.5

Anthropic

06 Oct 2026Safety AWAY 30

Anthropic expands verified cyber access through three security tiers

Anthropic combined Glasswing and its Cyber Verification Program into three access tiers. Verified defenders can request advanced Claude models; offensive testing requires authorization and extra controls. Its controlled Opus 5.5 evaluation blocked 46 of 50 Defense trials, while Red Team access completed 34 without blocks. These are test results, not real-world attacks.

Amazon 🇺🇸Anthropic 🇺🇸Google 🇺🇸Microsoft 🇺🇸Claude Fable 5.1 Claude Mythos 5.1 Claude Opus 5.5 Claude Sonnet 5.5

Anthropic

05 Oct 2026Governance AWAY 18

New York City Council hears AI companies and former staff on safety oversight

Representatives of OpenAI, Anthropic, Google and Meta testified at a New York City Council hearing examining AI risks and proposed oversight. Disclosure, independent review and whistleblower measures remained proposals, not enacted safeguards.

Anthropic 🇺🇸Google 🇺🇸Meta 🇺🇸OpenAI 🇺🇸

Associated Press

RELATED PEOPLE

People related to Google

Current and former roles are source-backed navigation metadata. A company relationship never changes an evidence assessment or the Doom Index.

Demis Hassabis 🇬🇧

Demis Hassabis is an AI researcher, co-founder of DeepMind and a leader at Google DeepMind whose public work spans general AI, scientific discovery and frontier-model governance.
- CEO, Google DeepMindcurrent · 2014 to presentSource
Jeff Dean 🇺🇸

Computer scientist and research leader whose work covers large-scale machine learning, frontier capabilities, deployment, and responsible AI.
- Chief Scientistcurrent · 1999 to presentSource
Rohin Shah 🇺🇸

Rohin Shah leads AGI Safety and Alignment at Google DeepMind and publishes research and first-person analysis on alignment, AI control, oversight, and frontier-model governance.
- AGI safety and alignment leadcurrent · Dates not establishedSource
- Head of AGI Safety and Alignmentcurrent · Dates not establishedSource
Victoria Krakovna 🇨🇦

AI-safety researcher whose work covers specification gaming, reward misspecification, agent incentives, and practical alignment methods.
- AI safety researchercurrent · Dates not establishedSource
Beth Barnes 🇺🇸

Founder and CEO of METR, where she leads independent evaluations of frontier-model autonomy, misalignment, and catastrophic risk. She previously conducted alignment and safety research at OpenAI and Google DeepMind.
- DeepMind researcherformer · Dates not establishedSource
Chris Olah 🇨🇦

Anthropic co-founder and interpretability researcher whose work develops methods for identifying features and circuits inside neural networks, with a focus on using mechanistic understanding to assess and improve advanced-AI safety.
- Researcher, Google Brainformer · Dates not establishedSource
Fei-Fei Li 🇨🇳

Computer scientist, Stanford professor and special adviser, and co-founder and CEO of World Labs, known for ImageNet, computer vision, spatial intelligence, and human-centered AI research and policy.
- Vice President and Chief Scientist of AI/ML, Google Cloudformer · From 2017Source
Geoffrey Hinton 🇬🇧

Computer scientist whose public commentary covers deep learning, advanced-AI capability, and catastrophic-risk concerns.
- Google Fellow and research adviserformer · 2013 to 2023Source
Ian Goodfellow 🇺🇸

Machine-learning researcher known for generative models and public technical work on robustness, security, and advanced-model behavior.
- Research scientistformer · 2017 to 2019Source

Videos about Google

Models from Google

Latest assessments involving Google

People related to Google

Share this page