Unverified50% confidenceBenchmarkExact time
HMA achieved an any-medal rate of 78.2% on the MLE-bench benchmark (covering 75 Kaggle tasks), setting a new state-of-the-art (SOTA).
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/12/2026
First Seen
Valid until: 1/10/2027
Sources
Related Entities
Related Claims
UnverifiedAcross six configurations covering five groups of heterogeneous agent combinations, HMA improved the any-medal rate by an average of 6.4 percentage points over its constituent baseline agents.67% similarUnverifiedDeepSeek V4.1 claims to achieve 74 on SWE, but the blogger found that this was the best result cherry-picked from multiple tests, with open code testing at 65 and PI testing at 66.61% similarUnverifiedBullet 在 SWE-bench Verified 基准测试中取得 95.8% 的成绩，位列前三，平均每个任务耗时 119 秒59% similarUnverifiedCognition has released SWE2, built on post-training of Kimi K3; the company claims it scores comparably to frontier models on major benchmarks but costs up to 70% less.59% similarUnverified在175张训练图/132张测试图的细粒度大众高尔夫车型分类任务中，SigLIP2 SO400M的k-NN准确率约为92%58% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1001126API
curl https://kongchang.com/api/v1/knowledge/claims/1001126MCP
get_claim(id=1001126)