AI Safety

What AI providers state their models reached in safety testing, in their own words, and what outside testers found.

Also: compare models side by side · statements that a capability level was crossed · providers · practices providers document · safety news.

What providers state

Each row is one statement in a provider's own safety document about a capability level of one of its models: the level as the provider names it, whether the provider says the model reached it, and the sentence that says so.

139 of 139

Model	Level, as the provider names it	The provider states	The provider's sentence
Muse Spark Meta	Moderate or lower Autonomy and AI research	Not reached	“We assess loss of control risk as “moderate or lower”.” source
Muse Spark Meta	High Chemical, biological, radiological, nuclear	Not determined	“Those evaluations identified elevated risks prior to mitigations, with Chemical and Biological capabilities assessed as likely reaching the “high risk” category under the Advanced AI Scaling Framework before safeguards were applied.” source
Muse Spark Meta	Moderate or lower Chemical, biological, radiological, nuclear	Not reached	“However, after applying mitigations, the residual risk of the model is that of “moderate or lower risk.”” source
Muse Spark Meta	Below the level required to meaningfully enable the cyber threat scenarios defin Cyber	Not reached	“Muse Spark’s offensive cyber capabilities remain below the level required to meaningfully enable the cyber threat scenarios defined in the Advanced AI Scaling Framework.” source
Muse Spark Meta	Moderate or lower Cyber	Not reached	“We assess cybersecurity risk as “moderate or lower” with respect to the threat scenarios outlined in Meta’s Advanced AI Scaling Framework.” source
Gemini 4 Argon Google DeepMind	TCL (Tracked Capability Level) Autonomy and AI research	Not determined	“In light of this, we assess that we cannot rule out the TCL.” source
Gemini 4 Argon Google DeepMind	TCL (Tracked Capability Level) Chemical, biological, radiological, nuclear	Not reached	“We can rule out the TCL for the chemical, radiological, and nuclear domains with reasonable confidence.” source
Gemini 4 Argon Google DeepMind	CCL (Critical Capability Level) Deception and sabotage	Not reached	“we assess its overall capabilities to be below the CCL threshold.” source
Gemini 4 Argon Google DeepMind	CCL (Critical Capability Level) Misuse and jailbreaks	Not reached	“Gemini 4 Argon demonstrates some ability to influence user beliefs and behaviors during one-on-one direct conversations in human behavioral studies, and may provide utility in the planning stage of a malicious influence operation but we assess its overall capabilities to be below the CCL threshold.” source
Gemini 4 Argon Google DeepMind	CCL (Critical Capability Level) Other	Not reached	“Gemini 4 Argon's performance remains below our CCL threshold.” source
Gemini 3.7 Flash Google DeepMind	CCL (Critical Capability Level) Autonomy and AI research	Not reached	“The model does not reach the CCL alert threshold.” source
Gemini 3.7 Flash Google DeepMind	CCL (Critical Capability Level) Chemical, biological, radiological, nuclear	Not reached	“We can rule out the CCL for the CBRN domain with reasonable conﬁdence based on the results from our testing.” source
Gemini 3.7 Flash Google DeepMind	TCL (Tracked Capability Level) Chemical, biological, radiological, nuclear	Not reached	“We can rule out the TCL for the CBRN domain with reasonable conﬁdence based on the results from our testing.” source
Gemini 3.7 Flash Google DeepMind	CCL (Critical Capability Level) Cyber	Not reached	“Gemini 3.7 Flash reaches the alert threshold for this CCL, but not the CCL.” source
Gemini 3.7 Flash Google DeepMind	TCL (Tracked Capability Level) Deception and sabotage	Not reached	“The model does not reach the TCL.” source
Gemini 3.5 Transcribe Live Google DeepMind	CCL (Critical Capability Level) Other	Not determined	“Our assessments have shown that none of Gemini 3.5 Translate, Gemini 3.5 Transcribe, or Gemini 3.5 Transcribe Live have meaningful new T/CCL-relevant capabilities or material increases in performance with respect to Frontier Safety compared to Gemini 3.1 Pro and Gemini 3.7 Flash, therefore we are conﬁdent that none of them are likely to reach any CCLs.” source
Gemini 3.5 Transcribe Google DeepMind	CCL (Critical Capability Level) Other	Not determined	“Our assessments have shown that none of Gemini 3.5 Translate, Gemini 3.5 Transcribe, or Gemini 3.5 Transcribe Live have meaningful new T/CCL-relevant capabilities or material increases in performance with respect to Frontier Safety compared to Gemini 3.1 Pro and Gemini 3.7 Flash, therefore we are conﬁdent that none of them are likely to reach any CCLs.” source
Gemini 3.5 Flash-Lite Google DeepMind	CCL (Critical Capability Level) Other	Not determined	“Our assessments have shown that, while Gemini 3.5 Flash-Lite excels at agents and coding, it does not have meaningful new capabilities or material increases in performance with respect to the domains outlined in our Frontier Safety Framework compared to Gemini 3.1 Pro; therefore, based on Gemini 3.1 Pro results, we are confident that Gemini 3.5 Flash-Lite is also unlikely to reach any CCLs.” source
Gemini 3.5 Audio Google DeepMind	Tracked or critical capability level (t/ccls) Other	Not reached	“they did not reach any Tracked or Critical Capability Levels (T/CCLs) outlined in our Frontier Safety Framework .” source
Gemini 3.1 Pro Google DeepMind	Uplift level 1 Chemical, biological, radiological, nuclear	Not reached	“We found that the model remains below alert thresholds for the CBRN, harmful manipulation, machine learning R&D, and misalignment CCLs.” source
Gemini 3.1 Pro Google DeepMind	Uplift level 1 Cyber	Not reached	“As previous models passed the alert threshold for cyber, we performed more additional testing in this domain on Gemini 3.1 Pro with and without Deep Think mode, and found that the model remains below the cyber CCL.” source
Gemini 3.1 Flash-Lite Google DeepMind	CCL (Critical Capability Level) Other	Not determined	“Our assessments have shown that Gemini 3.1 Flash-Lite is less capable than Gemini 3.1 Pro, therefore based on Gemini 3.1 Pro, we are confident that Gemini 3.1 Flash-Lite is also unlikely to reach any CCLs.” source
Gemini 3.1 Flash Audio Google DeepMind	CCL (Critical Capability Level) Other	Not determined	“Our assessments have shown that Gemini 3.1 Flash Audio is less capable than Gemini 3.1 Pro.” source
Gemini 2.5 Pro Google DeepMind	Machine Learning R&D Uplift Level 1 Autonomy and AI research	Not determined	“given Gemini 2.5 Pro Preview's poor median performance on RE-Bench relative to experts, the model likely lacks the necessary capabilities to automate or signiﬁcantly uplift any signiﬁcant fraction of the research process.” source
Gemini 2.5 Pro Google DeepMind	Instrumental Reasoning Level 1 Deception and sabotage	Not reached	“Instru- mental Reasoning Level 1 ✘ CCL not reached” source
Gemini 2.5 Pro Google DeepMind	Instrumental Reasoning Level 2 Deception and sabotage	Not reached	“Instru- mental Reasoning Level 2 ✘ CCL not reached” source
Gemini 2.5 Deep Think Google DeepMind	Uplift level 1 Chemical, biological, radiological, nuclear	Not determined	“Our assessment is that the model has enough technical knowledge in certain CBRN scenarios and stages to be considered at early warning.” source
Gemini 2.5 Deep Think Google DeepMind	Cyber Uplift Level 1 Cyber	Not determined	“CCL reached? No, but the early warning alert threshold was originally reached by Gemini 2.5 Pro and continues to be met by Gemini 2.5 Deep Think.” source
Gemini 2.5 Deep Think Google DeepMind	Uplift level 1 Cyber	Not determined	“For Cyber Uplift Level 1 , the early warning alert threshold was originally reached by Gemini 2.5 Pro and continues to be met by Gemini 2.5 Deep Think.” source
GPT-6.1 Terra OpenAI	High Autonomy and AI research	Not reached	“In AI Self-Improvement, GPT-6.1 Terra does not reach our High threshold.” source
GPT-6.1 Terra OpenAI	High Chemical, biological, radiological, nuclear	Reached	“Under our Preparedness Framework, we are treating GPT-6.1 Terra as Critical in cybersecurity and High for Biological and Chemical capability.” source
GPT-6.1 Terra OpenAI	Critical Cyber	Reached	“Under our Preparedness Framework, we are treating GPT-6.1 Terra as Critical in cybersecurity and High for Biological and Chemical capability.” source
GPT-5.5 Instant OpenAI	Below high capability Autonomy and AI research	Not reached	“GPT-5.5 Instant is considered below High Capability.” source
GPT-5.5 Instant OpenAI	High Autonomy and AI research	Not reached	“For AI Self-Improvement, evaluations of final checkpoints indicate that, like its predecessor models, GPT-5.5 Instant does not have a plausible chance of reaching a High threshold.” source
GPT-5.5 Instant OpenAI	High Chemical, biological, radiological, nuclear	Reached	“GPT-5.5 Instant is our first Instant model to be treated as High Capability in the Biological and Chemical domain, as well as in the Cybersecurity domain.” source
GPT-5.5 Instant OpenAI	High Cyber	Reached	“this is the first Instant model that we are treating as High capability in our Cybersecurity and Biological & Chemical Preparedness categories, and implementing appropriate safeguards.” source
GPT-5 OpenAI	High Chemical, biological, radiological, nuclear	Reached	“Similarly to ChatGPT agent, we have decided to treat gpt-5-thinking as High capability in the Biological and Chemical domain under our Preparedness Framework, activating the associated safeguards.” source
GPT-5 OpenAI	High cyber risk Cyber	Not reached	“The gpt-5 model series does not meet the threshold for high cyber risk.” source
Claude Sonnet 4.5 Anthropic	AI R&D-4 Autonomy and AI research	Not reached	“All models tested remain below our pre-committed rule-out threshold of 0.6.” source
Claude Sonnet 4.5 Anthropic	ASL-3 Autonomy and AI research	Not determined	“Exhibited improved autonomy capabilities in software engineering and AI research tasks, though still below the ASL-4 rule-out threshold.” source

Previous · Page 3 of 4 · Next

What outside testers report

Each row is one result from a government evaluator, independent lab, academic or security firm that tested a model, quoted from where it is reported.

1940 of 1940

Model	Tester	Test	What was found
Claude Mythos 5 Anthropic	Epoch AI Independent lab · 2026-06-11	cyber benchmarks Cyber	“The picture then changed after benchmarking the April version, as UK AISI did.” source
Claude Mythos 5 Anthropic	Epoch AI Independent lab · 2026-06-11	CyScenarioBench Cyber	“From Anthropic’s Mythos 5 launch post : Mythos 5: 36.7% Mythos Preview: 29.2% Opus 4.8: 16.6%” source
Claude Mythos 5 Anthropic	Epoch AI Independent lab · 2026-06-11	SCONE-Bench Cyber	“Anthropic created a benchmark of 405 Ethereum smart contracts that were exploited between 2020 and 2025.” source
Claude Mythos 5 Anthropic	Epoch AI Independent lab · 2026-06-11	OSS-Fuzz benchmark Cyber	“Mythos 5 triggered a crash 80% of the time, compared to 76.7% for Mythos Preview, and 61.5% for Opus 4.8.” source
Claude Mythos 5 Anthropic	Epoch AI Independent lab · 2026-06-11	UK AISI's CTF Suites Cyber	“AISI has a suite of capture-the-flag challenges where AI models must identify and exploit weaknesses in target systems to retrieve hidden “flags”.” source
Claude Mythos 5 Anthropic	Epoch AI Independent lab · 2026-06-11	UK AISI Cyber Ranges Cyber	“Although AISI reports that Mythos Preview (April) was able to fully complete “Cooling Tower” on 3/10 attempts (with all other models at 0/10), as they do not give its average performance we are not able to incorporate this information.” source
Claude Mythos 5 Anthropic	Epoch AI Independent lab · 2026-06-11	Microsoft CTI-REALM Cyber	“Given cyber threat intelligence reports, it tasks models to generate detection rules to apply on endpoint/cloud telemetry logs.” source
Claude Mythos 5 Anthropic	Epoch AI Independent lab · 2026-06-11	CVE-Bench Cyber	“CVE-Bench is a selection of 40 web application environments with known exploitable vulnerabilities in which models need to build exploits to achieve any one of 8 capabilities (access files, privilege escalation, etc.).” source
Claude Mythos 5 Anthropic	Epoch AI Independent lab · 2026-06-11	Cybench Cyber	“Cybench is a benchmark containing 40 professional-level Capture the Flag (CTF) tasks from 4 distinct CTF competitions, chosen to be recent (as of 2024), meaningful, and spanning a wide range of difficulties.” source
Claude Mythos 5 Anthropic	Epoch AI Independent lab · 2026-06-11	AISI’s Cyber Ranges Cyber	“A lot of the improvement comes from big jumps on ExploitGym, ExploitBench, AISI’s Cyber Ranges, and Anthropic’s SCONE-Bench.” source
GPT-5.5 OpenAI	Redwood Research Independent lab · 2026-06-10	no-CoT 50% time horizon (TH) on 43 benchmarks Other	“We find that frontier models like GPT-5.5 answer questions that take humans roughly three minutes with 50% reliability, and this TH has doubled approximately every year since 2019.” source
GPT-5.5 OpenAI	Redwood Research Independent lab · 2026-06-10	50% solve horizon (time horizon, TH) without chain-of-thought Autonomy and AI research	“GPT-5.5 is at the current frontier, with a 50% no-CoT TH of about 3 minutes and a reasoning token horizon of about 1,500 o3-mini tokens.” source
Claude Sonnet 4.6 Anthropic	SecureBio Independent lab · 2026-06-10	abc-bench fragment design Chemical, biological, radiological, nuclear	“on the fragment design task, models also demonstrated strong performance, with claude opus 4.6 achieving a perfect score across all runs.” source
Claude Sonnet 4.6 Anthropic	SecureBio Independent lab · 2026-06-10	abc-bench synthesis screening evasion Chemical, biological, radiological, nuclear	“in contrast, models performed worse on screening evasion, which has no published protocol and requires creative application of bioinformatics principles to a novel problem.” source
Claude Sonnet 4.6 Anthropic	SecureBio Independent lab · 2026-06-10	abc-bench liquid handling robot Chemical, biological, radiological, nuclear	“two models (claude sonnet 4.6 and gemini 3.1 pro preview) scored perfectly on the liquid handling robot task on all runs.” source
Claude Sonnet 4.6 Anthropic	SecureBio Independent lab · 2026-06-10	abc-bench fragment design task Chemical, biological, radiological, nuclear	“frontier models performed highly on both the fragment design and liquid handling robot tasks.” source
Claude Opus 4.7 Anthropic	Redwood Research Independent lab · 2026-06-10	43 benchmarks covering math, coding, knowledge, agentic tool-use, and safety-relevant questions like steganography and scheming reasoning Other	“We evaluated 14 frontier models from GPT-2 (2019) through GPT-5.5 (2026) on 43 benchmarks covering math, coding, knowledge, agentic tool-use, and safety-relevant questions like steganography and scheming reasoning.” source
Claude Opus 4.6 Anthropic	SecureBio Independent lab · 2026-06-10	abc-bench liquid handling robot Chemical, biological, radiological, nuclear	“frontier models performed highly on both the fragment design and liquid handling robot tasks.” source
Claude Mythos 5 Anthropic	Independent lab Independent lab · 2026-06-09	Black-box RNA sequence modeling and design challenge Chemical, biological, radiological, nuclear	“We partnered with Dyno Therapeutics on two sequence-to-function evaluations: a black-box RNA sequence modeling and design challenge benchmarked against 57 human participants drawn from the leading edge of the US ML-bio labor market, and an AAV capsid packaging prediction task measuring whether model domain knowledge and machine learning capabilities can outperform pretrained protein language models.” source
Claude Mythos 5 Anthropic	Frontier Design Independent lab · 2026-06-09	Internal and external red teaming with biological defense specialists Chemical, biological, radiological, nuclear	“To assess biologically-relevant capabilities across the full weapons development pipeline, we ran a portfolio of evaluations: internal and external (Deloitte) red teaming with biological defense specialists; the Frontier Design beneﬁcial red-teaming tabletop exercise pairing PhD biologists with LLM experts; a catastrophic biological scenario uplift trial (ﬁve three-person teams); and automated evaluations relevant to both CB-1 (long-form virology, VCT, DNA synthesis screening evasion) and CB-2 (black-box RNA sequence design and AAV capsid packaging prediction).” source
Claude Mythos 5 Anthropic	Independent lab Independent lab · 2026-06-09	Black-box RNA sequence design Chemical, biological, radiological, nuclear	“We partnered with Dyno Therapeutics on two evaluations of sequence-to-function modeling and design capability: 1. Black-box RNA sequence design: a medium-horizon challenge on which Dyno has evaluated 57 human participants drawn from the leading edge of the US ML-bio labor market since 2018.” source
Claude Mythos 5 Anthropic	Government evaluator Government evaluator · 2026-06-09	The Last Ones Cyber	“[Claude Mythos 5] matched Mythos Preview on "The Last Ones", solving the range end-to-end in 6/10 attempts.” source
Claude Mythos 5 Anthropic	Government evaluator Government evaluator · 2026-06-09	Doing Life Cyber	“No model we have tested has solved a new cyber range called "Doing Life", which is similar to “The Last Ones” but implements some basic cybersecurity defences.” source
Claude Mythos 5 Anthropic	Government evaluator Government evaluator · 2026-06-09	Cooling Tower Cyber	“[Claude Mythos 5] did not solve "Cooling Tower", an industrial control system range.” source
Claude Mythos 5 Anthropic	Government evaluator Government evaluator · 2026-06-09	expert-level cyber tasks Cyber	“On our expert-level cyber tasks, [Claude Mythos 5] performed comparably to Mythos Preview and ahead of GPT-5.5, the strongest fully public model tested.” source
Claude Mythos 5 Anthropic	Government evaluator Government evaluator · 2026-06-09	narrow cyber tasks Cyber	“We further assessed [Claude Mythos 5] against our narrow cyber tasks.” source
Claude Mythos 5 Anthropic	Independent lab Independent lab · 2026-06-09	red-teaming of classiﬁers on a ransomware-creation task by 10a Labs Cyber	“10a Labs spent about 20 hours red-teaming the classiﬁers on a ransomware-creation task using a variety of established jailbreaking techniques.” source
Claude Mythos 5 Anthropic	Independent lab Independent lab · 2026-06-09	red-teaming exercise by ALICE Cyber	“They found inconsistent blocking around marginal dual-use use-cases, but could not cause Opus 4.8 to complete any of the provided tasks.” source
Claude Mythos 5 Anthropic	Independent lab Independent lab · 2026-06-09	Shade Prompt injection and agents	“Mythos 5 had an attack success rate of 0.45% over all attempts, compared to 0.0% for Mythos Preview.” source
Claude Mythos 5 Anthropic	Government evaluator Government evaluator · 2026-06-09	third-party testing by the UK AI Security Institute Deception and sabotage	“We see this both in our own testing and in third-party testing by the UK AI Security Institute , though we do not ﬁnd it fully novel or actively dangerous.” source
Claude Mythos 5 Anthropic	UK AI Security Institute Government evaluator · 2026-06-09	evaluation awareness identification prompting Deception and sabotage	“When prompted to try to identify evaluations, Mythos 5 can do so in many cases across both the AI Security Institute’s testing and our own .” source
Claude Mythos 5 Anthropic	Government evaluator Government evaluator · 2026-06-09	open-ended testing of behaviors or risk factors related to misalignment Deception and sabotage	“[Claude Mythos 5] does not compromise safety research in our unprompted evaluations, does not fully refuse to assist with safety research, and has a reduced number of partial refusals to assist with safety research compared to the other models.” source
Claude Mythos 5 Anthropic	Independent lab Independent lab · 2026-06-09	Vending-Bench 2 Deception and sabotage	“Andon Labs tested the behavior of Claude Fable 5 16 in their simulated Vending-Bench 2 (as reported in Section 8.17.6 ) and Vending-Bench Arena retail-management evaluations.” source
Claude Mythos 5 Anthropic	Independent lab Independent lab · 2026-06-09	Vending-Bench Arena Deception and sabotage	“When competing head-to-head against Claude Opus 4.8 and GPT-5.5 in Vending-Bench Arena, Fable 5 was the only model to initiate price collusion.” source
Claude Mythos 5 Anthropic	Government evaluator Government evaluator · 2026-06-09	ControlARENA Deception and sabotage	“We used two evaluations: a sabotage capability evaluation (ControlARENA) and an evaluation of the model’s ability to control its chain-of-thought.” source
Claude Mythos 5 Anthropic	Government evaluator Government evaluator · 2026-06-09	CoT controllability evaluations from ‘Reasoning Models Struggle to Control their Chains of Thought’ Deception and sabotage	“We ran the CoT controllability evaluations from ‘Reasoning Models Struggle to Control their Chains of Thought’<link>, restricted to the GPQA, HLE and MMLU Pro datasets, with max effort across all models.” source
Claude Mythos 5 Anthropic	Independent lab Independent lab · 2026-06-09	CritPt Other	“We use the independent evaluation run by Artiﬁcial Analysis via the CritPt grading API.” source
Claude Fable 5 Anthropic	Independent lab Independent lab · 2026-06-09	METR external testing Autonomy and AI research	“External testing from AI safety researchers at METR was consistent with this conclusion.” source
Claude Fable 5 Anthropic	Independent lab Independent lab · 2026-06-09	external benchmark for prompt injection by Gray Swan Prompt injection and agents	“Notably, Mythos 5 obtained the lowest—that is, best—result yet seen on an external benchmark for prompt injection by Gray Swan.” source
Claude Fable 5 Anthropic	Independent lab Independent lab · 2026-06-09	external testing by METR Autonomy and AI research	“External testing by METR, described in Section 2.3.8 , reached conclusions consistent with this determination.” source

Previous · Page 31 of 49 · Next