Expired50% confidenceBenchmarkExact time
Hy3's score on the MRCR (Multi-turn Referential Coreference Resolution) benchmark jumped from 42.9% to 75.1%.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/11/2026
First Seen
This claim has expired. It is kept for the record and no longer describes the current state.
Valid until: 10/9/2026(expired)
Sources
Related Claims
UnverifiedOn the SWE-Bench Multilingual benchmark, HY3 scored 75.8, slightly behind DeepSeek V4 Pro's 76.2.66% similarUnverifiedHY3 still lags behind GLM 5.2 (the current top-level open-source model) in overall performance.64% similarUnverifiedOn SWE-Bench Pro, HY3 surpassed DeepSeek V4 Pro with a score of 57.9 versus 55.4.60% similarUnverifiedOmniMerge V4 achieved 73.4% on the MBPP code test, an improvement of 15.8 percentage points compared to the original Q6K quantization's 56.2%.59% similarUnverifiedIn evaluations, Qwen 3 Max outperformed Kimi K3 (77.5%), GLM (75%), and the GPT-4 series (71.25%).59% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/486977API
curl https://kongchang.com/api/v1/knowledge/claims/486977MCP
get_claim(id=486977)