Unverified50% confidenceFactExact time
The paper presents comparison data at a 10M token context on GLM-5.3-FLASH (QWEN-3.8-27B).
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
11/2/2026
First Seen
Valid until: 1/31/2027
Sources
Related Entities
Related Claims
UnverifiedIn a 10M token context test on GLM-5.3-FLASH (QWEN-3.8-27B), D-LSA achieves a prefill speedup of up to 72.4x and a decoding speedup of up to 42.5x compared with full attention.71% similarUnverifiedThe activated parameter count of Qwen3.8 Flash is approximately one-third that of GLM.67% similarUnverifiedThe model card gives GLM-5.3-Flash's benchmark scores: 84.3 on Terminal-Bench 2.1 and 63.4 on DeepSWE.67% similarUnverifiedGLM-5.3 has an input multiplier of 6.9, a cached input multiplier of 1.7, and an output multiplier of 24; GLM-5.3-Flash has an input multiplier of 2.3, a cached input multiplier of 0.56, and an output multiplier of 8.67% similarUnverifiedIn the GEPA paper, the merge operator can improve the average performance of gpt-4.1-mini but reduces the effectiveness of Qwen3-8B, an asymmetry that previously lacked explanation.65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1126142API
curl https://kongchang.com/api/v1/knowledge/claims/1126142MCP
get_claim(id=1126142)