BasedAGIBasedAGI
Menu
Rankings live

developer_tools

Debugging assistant

Localize bugs and propose fixes with explanations.

#1 Recommendation

gpt-4o-2024-05-13

Strong on RepoQA Official Results overall_average_pass_at_1_pct (99%) and RepoQA Official Results all_average_pass_at_1_pct (99%)

external/openai/gpt-4o-2024-05-13

21.6%

Score

26.3%

Confidence

Limited benchmark evidence for this use case.

54 ranked models with average evidence of 12.6 points. Rankings may shift as more benchmark data is ingested.

Ranked Models

30

Evidence Quality

81%

Scoring

Benchmark-backed

Top Signal

RepoQA Official Results: overall_average_pass_at_1_pct

All Ranked Models

Max params:
Min confidence:
30 of 30
RankModelScore
#2gpt-4o-2024-05-13

Strong on RepoQA Official Results overall_average_pass_at_1_pct (99%) and RepoQA Official Results all_average_pass_at_1_pct (99%)

21.6%
#3z-ai/glm-4.7

Strong on Sonar Java Quality Leaderboard functional_skill_pct (74%) and Vals LiveCodeBench overall_accuracy_pct (91%)

20.6%
#5Kimi K2 Thinking
20.1%
#6minimax/minimax-m2.1
19.2%
#10deepseek/deepseek-r1
17.2%
#11gemini-3-pro-preview
16.9%
#13google/gemini-3.1-pro-preview
16.5%
#14Grok-4-0709
16.3%
#15gpt-4.1-20250414
15.7%
#17openai/gpt-5.4-2026-03-05
15.5%
#19claude-sonnet-4-20250514
15.2%
#20anthropic/claude-sonnet-4.6
15.1%
#21anthropic/claude-opus-4-6-thinking
14.9%
#22Meta-Llama-3-70B-Instruct
14.9%
#23claude-opus-4-5-20251101
14.7%
#24GLM-5
14.6%
#25gemini-3-flash-preview
14.5%
#27gpt-5-2025-08-07
14.5%
#29gpt-5.2-2025-12-11
14.4%
#30gpt-5.1-2025-11-13
14.4%
#31anthropic/claude-opus-4-5-20251101-thinking
14.2%
#33gpt-4o
13.6%
#34kimi/kimi-k2.5-thinking
13.0%
#37Meta-Llama-3-8B-Instruct
12.6%
#39anthropic/claude-sonnet-4-5-20250929-thinking
12.5%
#40gpt-4o-2024-08-06
12.1%
#41gemini-2.5-pro
12.1%
#43zai/glm-5-thinking
11.6%
#44qwen-2.5-72b-instruct
11.6%
#47Phi-3-medium-128k-instruct
11.4%

Compare Models

Model A leads by +0.9%

Shareable Link →

Model A

gpt-4o-2024-05-13

external/openai/gpt-4o-2024-05-13

21.6%

Rank #2

Confidence 26.3%9 evidence pts

RepoQA Official Results: overall_average_pass_at_1_pct

Value 99.3% · Conf 100.0% · Weight 5.0%

repoqa_leaderboard.overall_average_pass_at_1_pct (Mar 12, 2026)

RepoQA Official Results: all_average_pass_at_1_pct

Value 99.3% · Conf 100.0% · Weight 1.7%

repoqa_leaderboard.all_average_pass_at_1_pct (Mar 12, 2026)

Aider Code Editing Leaderboard: percent_correct_pct

Value 82.3% · Conf 100.0% · Weight 1.3%

aider_code_editing.percent_correct_pct (Mar 12, 2026)

Aider Code Editing Leaderboard: correct_edit_format_pct

Value 85.7% · Conf 100.0% · Weight 0.7%

aider_code_editing.correct_edit_format_pct (Mar 12, 2026)

Model B

z-ai/glm-4.7

external/z-ai/glm-4-7

20.6%

Rank #3

Confidence 29.7%16 evidence pts

Sonar Java Quality Leaderboard: functional_skill_pct

Value 74.4% · Conf 100.0% · Weight 1.8%

sonar_java_quality.functional_skill_pct (Mar 12, 2026)

Vals LiveCodeBench: overall_accuracy_pct

Value 91.4% · Conf 100.0% · Weight 1.3%

vals_lcb.overall_accuracy_pct (Mar 12, 2026)

Vals SWE-bench: overall_accuracy_pct

Value 79.9% · Conf 100.0% · Weight 1.3%

vals_swebench.overall_accuracy_pct (Mar 12, 2026)

Sonar Java Quality Leaderboard: issue_density_error_per_kloc

Value 65.2% · Conf 100.0% · Weight 1.3%

sonar_java_quality.issue_density_error_per_kloc (Mar 12, 2026)

Ranking Diagnostics & Missing Models

Source Lift

Ranked

54

Sources

8

Quality

Insufficient

Vals LiveCodeBench

vals_lcb

39 rows

1.1% avg lift

Vals Legal Bench

vals_legal_bench

39 rows

0.3% avg lift

Vals Tax Eval v2

vals_tax_eval_v2

38 rows

0.2% avg lift

Vals CorpFin v2

vals_corp_fin_v2

38 rows

0.2% avg lift

Missing Strong Models

google/gemini-2.0-flash-001

external/google/gemini-2-0-flash-001

Rank #56

10.3%

Thin evidence after weighting

deepseek-v3

external/deepseek-ai/deepseek-v3

Rank #66

8.8%

Thin evidence after weighting

Llama-4-Scout-17B-16E-Instruct

external/meta/llama-4-scout-17b-16e-instruct

Rank #73

8.4%

Thin evidence after weighting
Taxonomy Details

Core Tasks

task.debug_localize_bugtask.debug_propose_fix

Required Modes

none

Domains

domain.software_engineering

Related Use Cases