Britain’s AI Security Institute has access to the latest advanced AI models and is evaluating them with the US government and major developers, including Google, AI Minister Kanishka Narayan said on Monday. The institute is testing models such as Gemini 4 Argon to identify risks early and assess safeguards, Narayan said. The announcement did not name all the models under evaluation.

The institute receives pre-deployment access under voluntary agreements with major AI developers, enabling it to assess models’ capabilities and risks and contribute to emerging international safety standards. Britain has no major AI developers of its own, and the government is seeking a leading role in assessing and managing risks from increasingly powerful AI systems. Microsoft President Brad Smith said in a BBC interview published on Monday that Britain was well placed to help lead work on AI safety guidelines.

The announcement follows heightened scrutiny after Anthropic and OpenAI disclosed cases of rogue or undesired behaviour by AI agents; some industry figures have warned that the technology could pose an existential threat to humanity. It also comes after reports last month that the White House asked OpenAI and Anthropic to withhold new models from British testers pending a US review. In a letter published in September, the AI Security Institute said Anthropic had stated that no organisation outside the US had access to its Mythos 5.1 model.

Asked about access to frontier AI systems, Narayan cited the UK’s long-standing relationship with the US and said he was confident the conversations between the governments would allow British researchers to test the latest generation of models. Ministers had never confirmed that the institute had broadly lost such access.

Britain’s AI Security Institute has access to the latest advanced AI models and is evaluating them with the US government and major developers, including Google, AI Minister Kanishka Narayan said on Monday. The institute is testing models such as Gemini 4 Argon to identify risks early and assess safeguards, Narayan said. The announcement did not name all the models under evaluation.

The institute receives pre-deployment access under voluntary agreements with major AI developers, enabling it to assess models’ capabilities and risks and contribute to emerging international safety standards. Britain has no major AI developers of its own, and the government is seeking a leading role in assessing and managing risks from increasingly powerful AI systems. Microsoft President Brad Smith said in a BBC interview published on Monday that Britain was well placed to help lead work on AI safety guidelines.

The announcement follows heightened scrutiny after Anthropic and OpenAI disclosed cases of rogue or undesired behaviour by AI agents; some industry figures have warned that the technology could pose an existential threat to humanity. It also comes after reports last month that the White House asked OpenAI and Anthropic to withhold new models from British testers pending a US review. In a letter published in September, the AI Security Institute said Anthropic had stated that no organisation outside the US had access to its Mythos 5.1 model.

Asked about access to frontier AI systems, Narayan cited the UK’s long-standing relationship with the US and said he was confident the conversations between the governments would allow British researchers to test the latest generation of models. Ministers had never confirmed that the institute had broadly lost such access.