Unverified50% confidenceBenchmarkExact time
Fabrication rates when the total is fully obscured: Claude Sonnet 5 at 38%, Gemini 3 Flash at 21%, Gemini 3.1 Flash-Lite at 44%, GPT-5.4 nano at 73%; fabrication rates when the total is half-obscured: Claude at 92%, the other three models all at 100%.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/11/2026
First Seen
Valid until: 1/9/2027
Sources
Related Entities
Related Claims
UnverifiedEven in the Indian context, 5% to 12% of responses from all models used Western thousands-separator formatting (Western comma drift) in intermediate steps, with Gemini 3.7 Flash the lowest (5%) and Claude Sonnet 4.6 the highest (12%).73% similarUnverifiedOn the Frontier Code programming benchmark, the old model scored 34.4%, while Gemini 3.7 Flash improved to 43.6%.68% similarUnverifiedIn the easy-to-verify category, seven models scored a perfect 40/40 across all six bait types, and Claude Sonnet 5.5, Gemini 3.7 Flash, and GPT-6.1 Sol answered every question correctly across all six tasks in both the easy and hard categories.67% similarUnverifiedOn the Frontier Code benchmark, GPT-6 Luna scored 56.1%, Claude Sonnet 5.0 scored 56.2%, and Kimi K3 scored 58.2%.66% similarUnverifiedWhen asked directly whether the total was visible, Claude, Gemini 3 Flash, and Flash-Lite answered across all 48 receipts that the total was not visible and could not be calculated, yet when asked directly for the total, they still fabricated one 21–44% of the time.66% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/997260API
curl https://kongchang.com/api/v1/knowledge/claims/997260MCP
get_claim(id=997260)