Inference
SWE-bench team

SWE-bench Multilingual · SWE-bench/experiments

Resolved-issue rate on issues across nine programming languages, from the submission records behind the leaderboard.

ModelsModels with a result.14
Top resultBest result on this test.N/A
Top-3 spreadPoints from first to third.N/A
UpdatedDate the source changed the results.2 Sept 2026

Result and cost

020406080100$0.03$0.1$0.3$1$3RESULTUS DOLLARS PER TASK · LOG SCALE

Results

14 results
#ModelResultThe score from the source.CostUS dollars to run one task.
—Gemini 3 Flashmini-SWE-agent72.7%$0.352
—Claude Opus 4.6mini-SWE-agent72.0%$0.663
—Claude Opus 4.5mini-SWE-agent70.7%$0.834
—GLM-5mini-SWE-agent69.7%$0.642
—Gemini 3 Promini-SWE-agent68.7%$1.02
—MiniMax 2.5mini-SWE-agent68.3%$0.101
—Kimi K2.5mini-SWE-agent67.3%$0.693
—Gemini 3.5 Flashmedium effort · mini-SWE-agent67.0%$0.627
—Claude Sonnet 4.5mini-SWE-agent67.0%$0.67
—GPT-5.2high effort · mini-SWE-agent66.7%$0.536
—GPT-5.2-Codexmini-SWE-agent66.3%$0.662
—Claude Haiku 4.5mini-SWE-agent64.7%$0.379
—DeepSeek V3.2mini-SWE-agent59.0%$0.384
—GPT-5 minimini-SWE-agent39.7%$0.052