Developer

Debugging assistant

Localize bugs and propose fixes with explanations.

task.debug_localize_bugtask.debug_propose_fix

Evidence quality is currently limited for this use case. Rankings below are useful for exploration, not a strong winner claim.

Provisional leader

claude-sonnet-4

Current leader based on limited benchmark evidence. Treat this ranking as directional until coverage improves.

29.8%

Best benchmark score

41.7%

Confidence

All ranked models — top 3

🥇

claude-sonnet-4

29.8%

🥈

gpt-4o-2024-05-13

27.2%

🥉

gpt-5-2025-08-07

26.0%

Ranked Models

Evidence Quality

81%

Evidence Points

Top Signal

SWE-bench Verified Leaderboard: swe_verified_resolved_pct

All Ranked Models

30 of 30 models

Rank	Model	Score	Confidence	Price / 1M	Evidence sources
🥇	claude-sonnet-4 Strong on SWE-bench Verified Leaderboard swe_verified_resolved_pct and Galileo Agent Leaderboard v2 Avg AC	29.8%	42%	$6.00	SWE-bench Verified LeaderboardGalileo Agent Leaderboard v2
🥈	gpt-4o-2024-05-13 Strong on RepoQA Official Results overall_average_pass_at_1_pct and SWE-bench Verified Leaderboard swe_verified_resolved_pct	27.2%	36%	—	RepoQA Official ResultsSWE-bench Verified Leaderboard
🥉	gpt-5-2025-08-07 Strong on SWE-bench Verified Leaderboard swe_verified_resolved_pct and Aider Polyglot Leaderboard percent_correct_pct	26.0%	31%	—	SWE-bench Verified LeaderboardAider Polyglot Leaderboard
#6	Kimi K2 Thinking Strong on SWE-bench Verified Leaderboard swe_verified_resolved_pct and Sonar Java Quality Leaderboard functional_skill_pct	23.3%	33%	$1.07	SWE-bench Verified LeaderboardSonar Java Quality Leaderboard
#8	gemini-3-pro-preview Strong on SWE-bench Verified Leaderboard swe_verified_resolved_pct and Vals SWE-bench overall_accuracy_pct	19.9%	24%	$4.50	SWE-bench Verified LeaderboardVals SWE-bench
#9	gemini-2.5-pro Strong on SWE-bench Verified Leaderboard swe_verified_resolved_pct and Galileo Agent Leaderboard v2 Avg AC	19.9%	29%	$3.44	SWE-bench Verified LeaderboardGalileo Agent Leaderboard v2
#10	gpt-5.2-2025-12-11 Strong on SWE-bench Verified Leaderboard swe_verified_resolved_pct and Vals SWE-bench overall_accuracy_pct	19.8%	23%	—	SWE-bench Verified LeaderboardVals SWE-bench
#12	o3-20250416 Strong on SWE-bench Verified Leaderboard swe_verified_resolved_pct and Aider Polyglot Leaderboard percent_correct_pct	18.9%	24%	$3.50	SWE-bench Verified LeaderboardAider Polyglot Leaderboard
#13	gpt-5-mini-2025-08-07 Strong on SWE-bench Verified Leaderboard swe_verified_resolved_pct and Vals LiveCodeBench overall_accuracy_pct	17.9%	26%	—	SWE-bench Verified LeaderboardVals LiveCodeBench
#14	glm-4.7 Strong on Sonar Java Quality Leaderboard functional_skill_pct and Vals LiveCodeBench overall_accuracy_pct	17.6%	26%	—	Sonar Java Quality LeaderboardVals LiveCodeBench
#15	deepseek-r1 Strong on Sonar Java Quality Leaderboard functional_skill_pct and Aider Polyglot Leaderboard percent_correct_pct	17.2%	25%	$0.27	Sonar Java Quality LeaderboardAider Polyglot Leaderboard
#17	claude-opus-4-5-20251101 Strong on SWE-bench Verified Leaderboard swe_verified_resolved_pct and Vals Terminal-Bench 2 overall_accuracy_pct	17.1%	20%	—	SWE-bench Verified LeaderboardVals Terminal-Bench 2
#18	gpt-4.1-20250414 Strong on Galileo Agent Leaderboard v2 Avg AC and SWE-bench Verified Leaderboard swe_verified_resolved_pct	16.9%	27%	—	Galileo Agent Leaderboard v2SWE-bench Verified Leaderboard
#20	minimax-m2.1 Strong on Sonar Java Quality Leaderboard functional_skill_pct and Vals SWE-bench overall_accuracy_pct	16.9%	26%	$0.53	Sonar Java Quality LeaderboardVals SWE-bench
#21	gpt-4o-20241120 Strong on Aider Code Editing Leaderboard percent_correct_pct and BigCodeBench Official bigcodebench_complete_pct	16.6%	30%	—	Aider Code Editing LeaderboardBigCodeBench Official
#22	gemini-3.1-pro-preview Strong on Vals SWE-bench overall_accuracy_pct and Vals Terminal-Bench 2 overall_accuracy_pct	16.1%	17%	$4.50	Vals SWE-benchVals Terminal-Bench 2
#23	claude-opus-4 Strong on SWE-bench Verified Leaderboard swe_verified_resolved_pct and Aider Polyglot Leaderboard percent_correct_pct	15.7%	19%	$10.00	SWE-bench Verified LeaderboardAider Polyglot Leaderboard
#25	o4-mini Strong on SWE-bench Verified Leaderboard swe_verified_resolved_pct and Aider Polyglot Leaderboard percent_correct_pct	15.1%	23%	$1.93	SWE-bench Verified LeaderboardAider Polyglot Leaderboard
#27	qwen-2.5-72b-instruct Strong on Galileo Agent Leaderboard v2 Avg AC and Aider Code Editing Leaderboard percent_correct_pct	14.6%	20%	—	Galileo Agent Leaderboard v2Aider Code Editing Leaderboard
#28	gpt-4.1 Strong on SWE-bench Verified Leaderboard swe_verified_resolved_pct and Aider Polyglot Leaderboard percent_correct_pct	14.5%	17%	$3.50	SWE-bench Verified LeaderboardAider Polyglot Leaderboard
#30	Grok-4-0709 Strong on Vals LiveCodeBench overall_accuracy_pct and Galileo Agent Leaderboard v2 Avg AC	14.2%	22%	—	Vals LiveCodeBenchGalileo Agent Leaderboard v2
#31	gpt-4o Strong on Sonar Java Quality Leaderboard functional_skill_pct and SWE-bench Verified Leaderboard swe_verified_resolved_pct	13.9%	30%	$0.26	Sonar Java Quality LeaderboardSWE-bench Verified Leaderboard
#34	gemini-3-flash-preview Strong on Vals SWE-bench overall_accuracy_pct and Vals LiveCodeBench overall_accuracy_pct	13.5%	16%	$1.13	Vals SWE-benchVals LiveCodeBench
#37	gpt-5.4-2026-03-05 Strong on Vals SWE-bench overall_accuracy_pct and Vals LiveCodeBench overall_accuracy_pct	13.5%	15%	—	Vals SWE-benchVals LiveCodeBench
#38	Phi-3-medium-128k-instruct Strong on RepoQA Official Results overall_average_pass_at_1_pct and RepoQA Official Results all_average_pass_at_1_pct	13.3%	19%	—	RepoQA Official ResultsRepoQA Official Results
#39	claude-sonnet-4.6 Strong on Vals SWE-bench overall_accuracy_pct and Vals LiveCodeBench overall_accuracy_pct	13.1%	15%	$6.00	Vals SWE-benchVals LiveCodeBench
#40	claude-opus-4-6-thinking Strong on Vals SWE-bench overall_accuracy_pct and Vals LiveCodeBench overall_accuracy_pct	12.9%	14%	—	Vals SWE-benchVals LiveCodeBench
#43	GLM-5 Strong on Sonar Java Quality Leaderboard functional_skill_pct and Sonar Java Quality Leaderboard issue_density_error_per_kloc	12.7%	14%	—	Sonar Java Quality LeaderboardSonar Java Quality Leaderboard
#45	gpt-5.1-2025-11-13 Strong on Vals LiveCodeBench overall_accuracy_pct and Vals SWE-bench overall_accuracy_pct	12.4%	15%	—	Vals LiveCodeBenchVals SWE-bench
#46	claude-opus-4-5-20251101-thinking Strong on Vals SWE-bench overall_accuracy_pct and Vals LiveCodeBench overall_accuracy_pct	12.4%	14%	—	Vals SWE-benchVals LiveCodeBench

Compare Models

Select two different models above to compare their evidence side by side.

▶Ranking diagnostics & missing models

Source lift

Ranked

Sources

Quality

Low

Vals LiveCodeBench

39 rows · 1.0% avg lift

Vals Legal Bench

39 rows · 0.2% avg lift

Vals Tax Eval v2

36 rows · 0.2% avg lift

Vals MedQA

34 rows · 0.2% avg lift

Missing frontier models

No obvious gaps right now.

▶Taxonomy & task details

Core tasks

task.debug_localize_bugtask.debug_propose_fix

Required modes

none

Domains

domain.software_engineering

Related in Developer

Codebase onboarding brief

Summarize a repository's architecture, modules, and conventions.

Autonomous Coding Agent

End-to-end autonomous software engineering: reading issues, writing code, running tests, submitting PRs.

Code generation

Generate correct, secure code from requirements.

Refactoring assistant

Refactor code safely while preserving behavior and improving clarity.