Unverified50% confidenceFactExact time

MAI Thinking-E在数学、科学等推理密集型领域表现较强，但在编码测试中仅达到DeepSeek V3.2的水平

1

Sources

50%

Confidence

Medium-term (~90 days)

Relevance

7/2/2026

First Seen

Valid until: 9/30/2026

Sources

微软Build 2026：自研推理模型MAI Thinking-E及AI全家桶深度解析

bilibiliinfinite灵感港6/3/2026

Related Claims

Unverified基准测试显示Qwen3-235B的表现追平甚至超越了DeepSeek R1，在数学、编码、推理等类别中均占优势70% similar UnverifiedPhi-3-mini（3.8B参数）在部分基准测试上超越了Mixtral 8x7B等更大的混合专家模型63% similar UnverifiedMAI Thinking-E总参数量约为1T（万亿），激活参数约35B，采用MoE（混合专家）架构62% similar UnverifiedQwen3.7 Max在数学推理、逻辑分析等高智力强度任务上保持了领先地位61% similar

Cite This Claim

Stable URI

https://kongchang.com/claim/46131

API

curl https://kongchang.com/api/v1/knowledge/claims/46131

MCP

get_claim(id=46131)