Unverified50% confidenceEventExact time
In a bug-fix test case, DeepSeek located and modified the calculation module so that all 8 tests passed, correcting the report from an erroneous $108.06 to the correct $65.05, taking approximately 1 minute and 20 seconds under max thinking.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/18/2026
First Seen
Valid until: 12/17/2026
Sources
Related Entities
Related Claims
UnverifiedIn the same test task, the DeepSeek V4 Pro preview version took approximately 5 minutes and 30 seconds but failed to fix the bug.67% similarUnverified初代 DeepSeek Math 7B 在 MATH 竞赛基准测试中以70亿参数达到 51.7% 的正确率63% similarUnverifiedGLM 5.2 scored 96 points in a real-world deep bug detection test, capable of identifying deep bugs nested more than three layers.59% similarUnverifiedOn the Aider Polyglot coding test, DeepSeek at full precision scored 71.6%, Unsloth's 4-bit dynamic version scored 69.7%, the 2-bit version scored 65.8%, while a competing quantization solution of the same 2-bit depth and file size scored only 56.6%.59% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/932810API
curl https://kongchang.com/api/v1/knowledge/claims/932810MCP
get_claim(id=932810)