Unverified60% confidenceFactTime unknown

前端生成测试中还对比了疑似Gemini 3的refer模型、Kimi KL和Haiku 4.5，GPT 5.1在指令遵循和视觉质感上处于领先

1

Sources

60%

Confidence

Medium-term (~90 days)

Relevance

6/3/2026

First Seen

Valid until: 9/1/2026

Sources

GPT 5.1 vs Claude Sonnet 4.5实测对比：推理、写作、编程谁更强

bilibilikate人不错

Related Entities

GPT 5.1model Gemini 3model Kimi KLmodel Haiku 4.5model

Related Claims

UnverifiedKimi K2.5在SWE（软件工程）测试中达到了Gemini 3 Pro的水平75% similar Unverified代码中还提到了一个名为'King4'的模型，据称在早期测试中表现出强劲性能，尚不清楚是Gemini 3的早期版本还是Gemini 2.5 Pro搭载Deep Think功能的变体72% similar VerifiedGemini系列从训练阶段就将多种模态联合建模，而非像GPT-4V等方案那样后接视觉编码器72% similar

Cite This Claim

Stable URI

https://kongchang.com/claim/40343

API

curl https://kongchang.com/api/v1/knowledge/claims/40343

MCP

get_claim(id=40343)