An Aleph Alpha study of models from Alibaba, DeepSeek, and Moonshot AI found that only 17% to 41% of responses on 967 hand-picked taboo topics, such as Tiananmen, Taiwan, and Xinjiang, were balanced

Overview Debates Related claims Topics Sources

Artificial Analysis reported that MiniMax M3.1 Flash demonstrated a 93 percent hallucination rate on the AA-Omniscience benchmark because the model attempted nearly every question rather than acknowledging uncertainty.

Artificial Analysis found MiniMax M3.1 Flash's Intelligence Index task evaluation cost dropped 57 percent to $0.22 from MiniMax M3's $0.51, compared to $0.26 for DeepSeek V4.1 Flash-1017, $0.28 for GLM-5.3 Flash, and $1.24 for Gemini 3.8 Flash.

MiniMax M3.1 Flash achieved an overall AA-Omniscience benchmark score of 16 despite correctly answering 56 percent of factual questions, compared to 43 percent for DeepSeek V4.1 Flash-1017 and 28 percent for GLM-5.3 Flash.

Artificial Analysis awarded MiniMax M3.1 Flash a score of 42 on its Intelligence Index, an increase of 13 points over the MiniMax M3 model's score of 29.

Artificial Analysis reported that DeepSeek V4.1 Flash-1017 scored 45 on its Intelligence Index, while Gemini 3.8 Flash scored 41.

MiniMax reported that all five reasoning effort levels for MiniMax M3.1 Flash stayed below $0.05 per task on its Automation-Bench chart.

Artificial Analysis reported that DeepSeek V4.1 Flash-1017 generated approximately 224 million output tokens during its Intelligence Index evaluation.

Artificial Analysis found MiniMax M3.1 Flash generated approximately 189 million output tokens during its Intelligence Index evaluation, compared to roughly 324 million for GLM-5.3 Flash and 334 million for Gemini 3.8 Flash.

Anthropic alleged in 2025 that state-sponsored Chinese hackers used its AI technology to automate break-ins targeting roughly 30 global targets, including corporations and foreign governments.

Japanese cybersecurity firm GMO Cybersecurity by Ierae reported on October 19, 2026, that around 80 Japanese companies across retail, insurance, logistics, and finance were targeted in cyberattacks using Chinese-developed AI agent Artex.

Frenos emphasized that its cyberattack tests used supplied information and simulated scenarios without proving that artificial intelligence models could independently identify targets or execute cyberattacks.

The Frenos study evaluated 10 artificial intelligence models across 16,980 total tests conducted between October 2025 and September 2026, including 13,884 decision attempts, 2,520 planning attempts, and 576 simulated attack runs.

The three leading frontier models tested by Frenos produced attack plans that passed every repeat test in 60%, 62.9%, and 74.3% of attempts, showing inconsistency and omitting essential steps or assuming unestablished access.

Single decision accuracy varied across models in the Frenos study, with models correctly matching required outcomes in 6% to 47% of single action choices.

An Anthropic report revealed that Chinese-linked operations utilized the Claude AI model to track Uyghurs and surveil religious and civil society groups.

Anthropic technical staff researcher Justin Curl stated that defining quality in the legal domain is notoriously difficult for artificial intelligence research

Anthropic offers four-month Conceptual Reasoning Fellowships paying between $100,000 and $167,000 to recruit legal experts to create evaluation test cases for its Claude AI models

A person close to Tencent stated that the scraping fleet originated from tests by a Tencent evaluation team checking how third-party models browse the web and answer questions.

Prof Tom Inglesby advocated for formal independent evaluations of new artificial intelligence models prior to release to ensure biological dangers are eliminated.

Anthropic is building its three-tier evaluation system where top-tier evaluators like METR receive employee-level access, while lower tiers like SecureBio run narrower reviews