Unverified70% confidenceBenchmarkExact time
On mainstream benchmarks such as MMLU, HumanEval, and GSM8K, open-source models have reached 85–95% of the level of early versions of GPT-4.
2
Sources
70%
Confidence
Medium-term (~90 days)
Relevance
8/13/2026
First Seen
Valid until: 11/11/2026
Sources
Related Claims
UnverifiedMMLU scores jumped from approximately 70% for GPT-3.5 to 86% for GPT-4, and then to above 90% for subsequent models.77% similarVerifiedGPT-5.5是OpenAI于2025年中期发布的模型，相比GPT-4o在推理深度和上下文理解方面有显著提升76% similarUnverifiedBy the end of 2024, open-source models such as Llama 3.1 405B and Qwen2.5-72B had approached or in some benchmarks even surpassed GPT-4 level performance.76% similarUnverifiedGPT-4在2023年初于MMLU上取得86.4%的成绩75% similarUnverifiedQwen2.5-72B approaches GPT-4 level performance on benchmarks such as MMLU and HumanEval.75% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/740520API
curl https://kongchang.com/api/v1/knowledge/claims/740520MCP
get_claim(id=740520)