AgentBench & ToolBench ↗4.01 points
gpt-4 / 0613 · AgentBench & ToolBench · AgentBench overall · Original standard test set ↗
OpenAI / MODEL EVIDENCE
GPT-4 / 0613
7 benchmarks · 12 observations · 6 sources
Scores at a glance
7 benchmarks · large dot: GPT-4 / 0613; small dots: closest models where reported
gpt-4-0613 · Aider coding benchmarks · Refactoring · percent correct ↗
HarmBench ↗21%
GPT-4-0613 · HarmBench · HarmBench harmful compliance · SORRY-Bench paper · cross-benchmark comparison ↗
GPT-4-0613 · StrongREJECT · StrongREJECT harmful compliance · SORRY-Bench paper · cross-benchmark comparison ↗
GPT-4 / 0613 · BigCodeBench & BigCodeBench-Hard · BigCodeBench Complete · greedy accuracy ↗
1 more reported result
45.9% · GPT-4 / 0613 · BigCodeBench & BigCodeBench-Hard · BigCodeBench Instruct · greedy accuracy · BigCodeBench: Benchmarking Code Generation with Diverse Function Calls and Complex Instructions ↗
GPT-4 (0613) · MMLU-Redux (2.0) · Virology · corrected-instance accuracy · MMLU-Redux 2.0 paper · audited subject comparison ↗
4 more reported results
99% · GPT-4 (0613) · MMLU-Redux (2.0) · Logical Fallacies · corrected-instance accuracy · MMLU-Redux 2.0 paper · audited subject comparison · Are We Done with MMLU? ↗
75% · GPT-4 (0613) · MMLU-Redux (2.0) · College Chemistry · corrected-instance accuracy · MMLU-Redux 2.0 paper · audited subject comparison · Are We Done with MMLU? ↗
68% · GPT-4 (0613) · MMLU-Redux (2.0) · Professional Law · corrected-instance accuracy · MMLU-Redux 2.0 paper · audited subject comparison · Are We Done with MMLU? ↗
43% · GPT-4 (0613) · MMLU-Redux (2.0) · Human Sexuality · corrected-instance accuracy · MMLU-Redux 2.0 paper · audited subject comparison · Are We Done with MMLU? ↗
ForecastBench ↗61.5 Brier index
gpt-4-0613† · Overall difficulty-adjusted Brier index ↗
Closest models by score
Similar positions on shared benchmark leaderboards. Smaller gaps are closer.
There are not enough shared scored settings to identify close comparisons reliably.
Sources
Every plotted observation links to its evaluation. Reported configurations, benchmark versions and task subsets can differ; similarity describes shared scores, not training lift.
- Aider refactoring benchmark ↗
- AgentBench: Evaluating LLMs as Agents — published result table ↗
All 22 authors
Xiao Liu, Hao Yu, Hanchen Zhang, Yifan Xu, Xuanyu Lei, Hanyu Lai, Yu Gu, Hangliang Ding, Kaiwen Men, Kejuan Yang, Shudan Zhang, Xiang Deng, Aohan Zeng, Zhengxiao Du, Chenhui Zhang, Sheng Shen, Tianjun Zhang, Yu Su, Huan Sun, Minlie Huang, Yuxiao Dong, Jie Tang
- Are We Done with MMLU? — published result table ↗
All 16 authors
Gema, Aryo Pradipta, Leang, Joshua Ong Jun, Hong, Giwon, Devoto, Alessio, Mancino, Alberto Carlo Maria, Saxena, Rohit, He, Xuanli, Zhao, Yu, Du, Xiaotang, Madani, Mohammad Reza Ghasemi, Barale, Claire, McHardy, Robert, Harris, Joshua, Kaddour, Jean, van Krieken, Emile, Minervini, Pasquale
- SORRY-Bench: Systematically Evaluating Large Language Model Safety Refusal — published result table ↗
All 16 authors
Tinghao Xie, Xiangyu Qi, Yi Zeng, Yangsibo Huang, Udari Madhushani Sehwag, Kaixuan Huang, Luxi He, Boyi Wei, Dacheng Li, Ying Sheng, Ruoxi Jia, Bo Li, Kai Li, Danqi Chen, Peter Henderson, Prateek Mittal
- BigCodeBench: Benchmarking Code Generation with Diverse Function Calls and Complex Instructions — published result table ↗
All 33 authors
Terry Yue Zhuo, Minh Chien Vu, Jenny Chim, Han Hu, Wenhao Yu, Ratnadira Widyasari, Imam Nur Bani Yusuf, Haolan Zhan, Junda He, Indraneil Paul, Simon Brunner, Chen Gong, Thong Hoang, Armel Randy Zebaze, Xiaoheng Hong, Wen-Ding Li, Jean Kaddour, Ming Xu, Zhihan Zhang, Prateek Yadav, Naman Jain, Alex Gu, Zhoujun Cheng, Jiawei Liu, Qian Liu, Zijian Wang, Binyuan Hui, Niklas Muennighoff, David Lo, Daniel Fried, Xiaoning Du, Harm de Vries, Leandro Von Werra
- Forecastbench — methodology and published results ↗
Names used in source tables
GPT-4 (0613) · GPT-4 / 0613 · GPT-4-0613 · gpt-4 / 0613 · gpt-4-0613 · gpt-4-0613†