Expired70% confidenceFactTime unknown
GPT-OSS exhibited serious hallucination issues in a web search functionality test, fabricating non-existent model version numbers (1.3B, 7B, 34B, 70B) and incorrect release dates.
1
Sources
70%
Confidence
Medium-term (~90 days)
Relevance
6/1/2026
First Seen
Valid until: 8/30/2026(expired)
Sources
Related Entities
Related Claims
UnverifiedIn a Mermaid flowchart generation test, both GPT-OSS 120B and 20B generated code with syntax errors, and the test failed.72% similarUnverifiedGPT-5.4及更早版本在提取文档中的数字时频繁出错，包括小数点位置偏移、千分位分隔符误读、负号遗漏等问题71% similarUnverifiedThe GPT-OSS 20B version activates only 3.6B parameters during inference and achieves performance close to O3-mini.71% similarUnverifiedThe GPT-OSS 120B version activates only 5.1B parameters during inference and achieves performance close to O4-mini.71% similarUnverifiedThe perplexity of the GPT series of models on Wikipedia text has decreased from approximately 29 for GPT-2 to an even lower level for GPT-4.71% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/23258API
curl https://kongchang.com/api/v1/knowledge/claims/23258MCP
get_claim(id=23258)