Models / OpenAI

OpenAI / MODEL EVIDENCE

GPT-4-0125-preview

6 benchmarks · 67 observations · 5 sources

Scores at a glance

6 benchmarks · large dot: GPT-4-0125-preview; small dots: closest models where reported

Aider coding benchmarks ↗66.2%

gpt-4-0125-preview · Aider coding benchmarks · Refactoring · percent correct ↗

CFBench (Chinese Complex Following) ↗84.3%

GPT-4-0125-Preview† · CFBench (Chinese Complex Following) · Full set · Constraint satisfaction · full-set evaluation ↗

2 more reported results

55% · GPT-4-0125-Preview† · CFBench (Chinese Complex Following) · Full set · Instruction satisfaction · full-set evaluation · CFBench: A Comprehensive Constraints-Following Benchmark for LLMs ↗

64.7% · GPT-4-0125-Preview† · CFBench (Chinese Complex Following) · Full set · Priority satisfaction · full-set evaluation · CFBench: A Comprehensive Constraints-Following Benchmark for LLMs ↗

HumanEval ↗86.6%

gpt-4-0125-preview · HumanEval · Pass@1 · OpenAI Simple Evals · archived July 2025 ↗

DROP (Discrete Reasoning Over Paragraphs) ↗81.5%

gpt-4-0125-preview · DROP (Discrete Reasoning Over Paragraphs) · F1 · three-shot · OpenAI Simple Evals · archived July 2025 ↗

NYU CTF Bench ↗5%

gpt-4-0125-preview · NYU CTF Bench · pass@5 · 200-challenge full set ↗

6 more reported results

0% · gpt-4-0125-preview · NYU CTF Bench · pass@5 · crypto subset · NYU CTF Bench: A Scalable Open-Source Benchmark Dataset for Evaluating LLMs in Offensive Security — official leaderboard ↗

6.6667% · gpt-4-0125-preview · NYU CTF Bench · pass@5 · forensics subset · NYU CTF Bench: A Scalable Open-Source Benchmark Dataset for Evaluating LLMs in Offensive Security — official leaderboard ↗

0% · gpt-4-0125-preview · NYU CTF Bench · pass@5 · misc subset · NYU CTF Bench: A Scalable Open-Source Benchmark Dataset for Evaluating LLMs in Offensive Security — official leaderboard ↗

7.6923% · gpt-4-0125-preview · NYU CTF Bench · pass@5 · pwn subset · NYU CTF Bench: A Scalable Open-Source Benchmark Dataset for Evaluating LLMs in Offensive Security — official leaderboard ↗

9.8039% · gpt-4-0125-preview · NYU CTF Bench · pass@5 · rev subset · NYU CTF Bench: A Scalable Open-Source Benchmark Dataset for Evaluating LLMs in Offensive Security — official leaderboard ↗

5.2632% · gpt-4-0125-preview · NYU CTF Bench · pass@5 · web subset · NYU CTF Bench: A Scalable Open-Source Benchmark Dataset for Evaluating LLMs in Offensive Security — official leaderboard ↗

BABILong ↗87%

gpt-4-0125-preview · BABILong · BABILong v0 · QA1–QA5 average · 0k context · historical v0 leaderboard ↗

53 more reported results

74% · gpt-4-0125-preview · BABILong · BABILong v0 · QA1–QA5 average · 4k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

64% · gpt-4-0125-preview · BABILong · BABILong v0 · QA1–QA5 average · 16k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

43% · gpt-4-0125-preview · BABILong · BABILong v0 · QA1–QA5 average · 64k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

36% · gpt-4-0125-preview · BABILong · BABILong v0 · QA1–QA5 average · 128k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

81% · gpt-4-0125-preview · BABILong · BABILong v0 · QA1–QA5 average · 1k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

77% · gpt-4-0125-preview · BABILong · BABILong v0 · QA1–QA5 average · 2k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

71% · gpt-4-0125-preview · BABILong · BABILong v0 · QA1–QA5 average · 8k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

53% · gpt-4-0125-preview · BABILong · BABILong v0 · QA1–QA5 average · 32k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

99% · gpt-4-0125-preview · BABILong · BABILong v0 · QA1 · 0k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

100% · gpt-4-0125-preview · BABILong · BABILong v0 · QA1 · 1k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

100% · gpt-4-0125-preview · BABILong · BABILong v0 · QA1 · 2k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

95% · gpt-4-0125-preview · BABILong · BABILong v0 · QA1 · 4k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

93% · gpt-4-0125-preview · BABILong · BABILong v0 · QA1 · 8k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

84% · gpt-4-0125-preview · BABILong · BABILong v0 · QA1 · 16k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

62% · gpt-4-0125-preview · BABILong · BABILong v0 · QA1 · 32k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

58% · gpt-4-0125-preview · BABILong · BABILong v0 · QA1 · 64k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

39% · gpt-4-0125-preview · BABILong · BABILong v0 · QA1 · 128k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

88% · gpt-4-0125-preview · BABILong · BABILong v0 · QA2 · 0k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

79% · gpt-4-0125-preview · BABILong · BABILong v0 · QA2 · 1k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

72% · gpt-4-0125-preview · BABILong · BABILong v0 · QA2 · 2k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

68% · gpt-4-0125-preview · BABILong · BABILong v0 · QA2 · 4k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

65% · gpt-4-0125-preview · BABILong · BABILong v0 · QA2 · 8k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

59% · gpt-4-0125-preview · BABILong · BABILong v0 · QA2 · 16k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

42% · gpt-4-0125-preview · BABILong · BABILong v0 · QA2 · 32k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

25% · gpt-4-0125-preview · BABILong · BABILong v0 · QA2 · 64k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

25% · gpt-4-0125-preview · BABILong · BABILong v0 · QA2 · 128k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

56% · gpt-4-0125-preview · BABILong · BABILong v0 · QA3 · 0k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

63% · gpt-4-0125-preview · BABILong · BABILong v0 · QA3 · 1k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

57% · gpt-4-0125-preview · BABILong · BABILong v0 · QA3 · 2k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

56% · gpt-4-0125-preview · BABILong · BABILong v0 · QA3 · 4k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

53% · gpt-4-0125-preview · BABILong · BABILong v0 · QA3 · 8k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

45% · gpt-4-0125-preview · BABILong · BABILong v0 · QA3 · 16k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

31% · gpt-4-0125-preview · BABILong · BABILong v0 · QA3 · 32k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

31% · gpt-4-0125-preview · BABILong · BABILong v0 · QA3 · 64k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

32% · gpt-4-0125-preview · BABILong · BABILong v0 · QA3 · 128k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

98% · gpt-4-0125-preview · BABILong · BABILong v0 · QA4 · 0k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

70% · gpt-4-0125-preview · BABILong · BABILong v0 · QA4 · 1k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

63% · gpt-4-0125-preview · BABILong · BABILong v0 · QA4 · 2k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

60% · gpt-4-0125-preview · BABILong · BABILong v0 · QA4 · 4k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

52% · gpt-4-0125-preview · BABILong · BABILong v0 · QA4 · 8k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

47% · gpt-4-0125-preview · BABILong · BABILong v0 · QA4 · 16k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

46% · gpt-4-0125-preview · BABILong · BABILong v0 · QA4 · 32k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

40% · gpt-4-0125-preview · BABILong · BABILong v0 · QA4 · 64k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

32% · gpt-4-0125-preview · BABILong · BABILong v0 · QA4 · 128k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

96% · gpt-4-0125-preview · BABILong · BABILong v0 · QA5 · 0k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

95% · gpt-4-0125-preview · BABILong · BABILong v0 · QA5 · 1k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

92% · gpt-4-0125-preview · BABILong · BABILong v0 · QA5 · 2k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

90% · gpt-4-0125-preview · BABILong · BABILong v0 · QA5 · 4k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

93% · gpt-4-0125-preview · BABILong · BABILong v0 · QA5 · 8k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

85% · gpt-4-0125-preview · BABILong · BABILong v0 · QA5 · 16k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

82% · gpt-4-0125-preview · BABILong · BABILong v0 · QA5 · 32k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

60% · gpt-4-0125-preview · BABILong · BABILong v0 · QA5 · 64k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

51% · gpt-4-0125-preview · BABILong · BABILong v0 · QA5 · 128k context · historical v0 leaderboard · BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard ↗

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.

Filter comparisons to this model →

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 ↗
Paul Gauthier, Aider contributors
CFBench: A Comprehensive Constraints-Following Benchmark for LLMs — published result table ↗
Tao Zhang, Chenglin Zhu, Yanjun Shen, Wenjing Luo, Yan Zhang, Hao Liang et al. · 2024
All 14 authors
Tao Zhang, Chenglin Zhu, Yanjun Shen, Wenjing Luo, Yan Zhang, Hao Liang, Tao Zhang, Fan Yang, Mingan Lin, Yujing Qiao, Weipeng Chen, Bin Cui, Wentao Zhang, Zenan Zhou
BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystack — official v0 leaderboard — published result table ↗
Kuratov, Yuri, Bulatov, Aydar, Anokhin, Petr, Rodkin, Ivan, Sorokin, Dmitry, Sorokin, Artyom et al. · 2024
All 7 authors
Kuratov, Yuri, Bulatov, Aydar, Anokhin, Petr, Rodkin, Ivan, Sorokin, Dmitry, Sorokin, Artyom, Burtsev, Mikhail
NYU CTF Bench: A Scalable Open-Source Benchmark Dataset for Evaluating LLMs in Offensive Security — official leaderboard ↗
Minghao Shao, Sofija Jancheska, Meet Udeshi, Brendan Dolan-Gavitt, Haoran Xi, Kimberly Milner et al. · 2024
All 13 authors
Minghao Shao, Sofija Jancheska, Meet Udeshi, Brendan Dolan-Gavitt, Haoran Xi, Kimberly Milner, Boyuan Chen, Max Yin, Siddharth Garg, Prashanth Krishnamurthy, Farshad Khorrami, Ramesh Karri, Muhammad Shafique
Simple Evals — model benchmark results ↗
OpenAI · 2025

Names used in source tables

GPT-4-0125-Preview† · gpt-4-0125-preview