Unverified50% confidenceBenchmarkExact time
在HLE科学类基准上，Agent A1 据实测数据超过了Qwen 3.6和Step 3.5 Flash
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/22/2026
First Seen
Valid until: 10/20/2026
Sources
Related Claims
UnverifiedStep 3.7 Flash 在 ClawEval-1.1 基准测试中以 67.1 分领先71% similarUnverified在本地推理速度对比中，Agent A1（95 tokens/秒）快于Gemma 4 Coder和Qwen系列67% similarUnverifiedStep 3.7 Flash 在 SimpleVQA 基准测试中以 79.2 分排名第一65% similarUnverifiedThe official statement indicates that HY3 outperforms Deepseek V4 Flash, which has a comparable parameter scale, in most scenarios, and scored higher on average than the GLM series models in internal blind tests64% similarUnverifiedPrime Agent, paired with the Opus 5 model, achieved a score of 95.5% on the ARC-AGI-3 benchmark, reportedly surpassing the human expert baseline reported in the paper.60% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/593887API
curl https://kongchang.com/api/v1/knowledge/claims/593887MCP
get_claim(id=593887)