Unverified50% confidenceBenchmarkExact time
V4 Pro Max满血推理模式下算法竞赛3206分、编程93.5分、科学问答90.1分、软件工程80.6分、终端操作67.9分
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
8/19/2026
First Seen
Valid until: 11/17/2026
Sources
Related Claims
UnverifiedThis test selected five AI models — DeepSeek R1, Claude Sonnet 3.7, ChatGPT o3 Mini, Grok 3, and Tongyi Qianwen 2.5 Max — and used a unified prompt to have each of them generate a runnable web-based Snake game.65% similarUnverified推理端优化技术包括量化、知识蒸馏和KV Cache，量化可将参数从32位或16位浮点压缩为8位甚至4位整数64% similarUnverified叠加cv=5的5折交叉验证后，144种参数组合需训练144×5=720次模型64% similarUnverified根据Cursor自研的CursorBench基准测试，Opus 4.8相比前代Opus 4.7在编码效率上有显著提升63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/774252API
curl https://kongchang.com/api/v1/knowledge/claims/774252MCP
get_claim(id=774252)