Unverified50% confidenceBenchmarkExact time
On the math reasoning task GSM8K, OQ4E answered 954 questions correctly (95.4%), while 8-bit only answered 924 correctly (92.4%), with the 4-bit version coming out 3 percentage points higher.
1
Sources
50%
Confidence
Long-term
Relevance
9/11/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedOn the code generation task HumanEval, 8-bit scored 92.7%, significantly higher than OQ4E's 89.6%.75% similarUnverifiedIn the MMLU test, OQ4E surpassed 8-bit with 84.5% versus 82.8%.70% similarVerifiedGGUF量化格式（如Q4_K_M）通过将权重精度从32位浮点压缩到4位整数，通常只带来约1-5%的性能损失69% similarUnverifiedQuantization methods such as GPTQ, AWQ, and SqueezeLLM can still maintain over 95% of the original model's performance at 4-bit precision.68% similarUnverifiedV4 Pro Max满血推理模式下算法竞赛3206分、编程93.5分、科学问答90.1分、软件工程80.6分、终端操作67.9分68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/907289API
curl https://kongchang.com/api/v1/knowledge/claims/907289MCP
get_claim(id=907289)