Unverified50% confidenceFactExact time
In the honesty test, the number of confident wrong answers was: Opus one, Sonnet four, DeepSeek Flash eighteen, and DeepSeek Pro twenty-six.
1
Sources
50%
Confidence
Short-term (~14 days)
Relevance
10/21/2026
First Seen
Valid until: 11/4/2026
Sources
Related Entities
Related Claims
UnverifiedIn the new version of the 24-point problem (numbers 2, 7, 8, 10, requiring three different solutions), DeepSeek V4.1 Flash got all three solutions wrong, each violating the constraint that each number can only be used once.73% similarUnverifiedIn a test of 360 clean texts, Jev made only 1 misjudgment, gpt-5-mini made 1 misjudgment, Gemini 2.5 Flash-Lite and Gemma each made 2 misjudgments, gpt-oss-safeguard-20b made 4 misjudgments, and Llama Guard 4 made 26 misjudgments.68% similarUnverifiedIn the task of completing the last Token of an FP8 quantization function, after DeepSeek-V4-Flash-Base adds equal-sign decorations before the code, when the number of equal signs modulo 4 is 0 or 1 the error rate is as high as 71.3%, and when the remainder is 2 or 3 the accuracy returns to 91.5%.67% similarUnverifiedDeepSeek V4.1 claims to achieve 74 on SWE, but the blogger found that this was the best result cherry-picked from multiple tests, with open code testing at 65 and PI testing at 66.66% similarUnverifiedCaveman's fixed benchmark test results from 54 runs show a 33.2% reduction in input tokens, with all 18/18 correctness checks passing.66% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1081796API
curl https://kongchang.com/api/v1/knowledge/claims/1081796MCP
get_claim(id=1081796)