US Model Opus 4.8 Accused of Distilling Chinese Open-Source AI
Anthropic, a leading U.S. large language model company, quietly released the Claude Opus 4.8 model at midnight today. While it achieved strong results across several benchmarks, many developers and netizens reported that the model appears to suffer from "cognitive confusion."
During an API test, one user found that when asked about its own identity, the model incorrectly claimed to be either Alibaba's "Qwen" or "DeepSeek," two well-known open-source Chinese models.
API Test Triggers Identity Confusion
Some users testing through the web interface said they could not replicate the issue and argued that the model had not been distilled from Chinese models. However, developers familiar with the matter explained that the web interface uses strict system prompts, which hide this behavior.
In contrast, when tested through a minimal API setup with no system prompts, the issue appeared frequently. This strongly suggests that the top-tier U.S. model may have been trained or fine-tuned using a large amount of generated data from Chinese open-source models through a process of distillation.
Inconsistent Policies Draw Accusations of Double Standards
The incident stirred up considerable discussion in the tech community, largely because Anthropic had previously taken a strong stance against China. The company worked with the U.S. Department of Defense and other agencies to draft legislation that sought to classify large model distillation as an adversarial threat.
They not only urged the U.S. government to tighten chip and software restrictions on China's AI industry, but also set a strategic goal of maintaining a two-year lead over China in AI by 2028. Now, with their flagship model allegedly caught distilling from Chinese models, many see an ironic case of double standards.
Related article
DeepMind CEO Hassabis: I sleep six hours a day, usually feel energetic around 1 a.m.
Fortune recently featured an interview with Demis Hassabis, CEO of Google DeepMind, revealing his unconventional approach to rest and productivity. Hassabis disclosed that he sleeps very little, structuring his waking hours into two distinct work blo
OpenAI, Anthropic Vie for Market Share Despite Revenue Shortfalls
Despite recent reports suggesting OpenAI missed revenue targets, creating pressure on tech stocks this Tuesday, private AI lab investors remain resilient. Seasoned backers have confirmed they will not reduce investment despite negative media coverage
California AV Compliance: A New Era of Tickets, Geofences, and 1M Miles
Guident operates an AuveTech shuttle in South Florida, managing a four-mile route in West Palm Beach and a one-mile route in Boca Raton using its remote monitoring technology. | Credit: GuidentCalifornia is redefining the regulatory landscape for dri
Related Special Topic Recommendations
Comments (1)
0/500
Anthropic, a leading U.S. large language model company, quietly released the Claude Opus 4.8 model at midnight today. While it achieved strong results across several benchmarks, many developers and netizens reported that the model appears to suffer from "cognitive confusion."
During an API test, one user found that when asked about its own identity, the model incorrectly claimed to be either Alibaba's "Qwen" or "DeepSeek," two well-known open-source Chinese models.
API Test Triggers Identity Confusion
Some users testing through the web interface said they could not replicate the issue and argued that the model had not been distilled from Chinese models. However, developers familiar with the matter explained that the web interface uses strict system prompts, which hide this behavior.
In contrast, when tested through a minimal API setup with no system prompts, the issue appeared frequently. This strongly suggests that the top-tier U.S. model may have been trained or fine-tuned using a large amount of generated data from Chinese open-source models through a process of distillation.
Inconsistent Policies Draw Accusations of Double Standards
The incident stirred up considerable discussion in the tech community, largely because Anthropic had previously taken a strong stance against China. The company worked with the U.S. Department of Defense and other agencies to draft legislation that sought to classify large model distillation as an adversarial threat.
They not only urged the U.S. government to tighten chip and software restrictions on China's AI industry, but also set a strategic goal of maintaining a two-year lead over China in AI by 2028. Now, with their flagship model allegedly caught distilling from Chinese models, many see an ironic case of double standards.
DeepMind CEO Hassabis: I sleep six hours a day, usually feel energetic around 1 a.m.
Fortune recently featured an interview with Demis Hassabis, CEO of Google DeepMind, revealing his unconventional approach to rest and productivity. Hassabis disclosed that he sleeps very little, structuring his waking hours into two distinct work blo
OpenAI, Anthropic Vie for Market Share Despite Revenue Shortfalls
Despite recent reports suggesting OpenAI missed revenue targets, creating pressure on tech stocks this Tuesday, private AI lab investors remain resilient. Seasoned backers have confirmed they will not reduce investment despite negative media coverage





Home






