Expired50% confidenceFactExact time
DGM automatically more than doubled baseline performance on the SWE-bench software engineering benchmark, achieving an absolute improvement of 30 percentage points
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/2/2026
First Seen
This claim has expired. It is kept for the record and no longer describes the current state.
Valid until: 9/30/2026(expired)
Sources
Sakana AI Establishes Recursive Self-Improvement Lab, Replacing Brute-Force Compute with Creativity to Pioneer a New AI Paradigm
rss6/4/2026
Related Claims
UnverifiedArtificial Analysis给Muse Glimmer 30B打出35分的智能指数，相比Llama 4 Maverick是巨大飞跃64% similarUnverified基于尾调用的字节码分发在某些基准测试中可带来 5-15% 的性能提升64% similarUnverifiedIn the temperature control task, the learned policy improved from -304 to -35.8, demonstrating a significant advantage.62% similarUnverifiedApple Silicon Neural Engine的峰值算力从M1的11 TOPS跃升至M4的38 TOPS，性能提升超过3倍60% 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60% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/42235API
curl https://kongchang.com/api/v1/knowledge/claims/42235MCP
get_claim(id=42235)