AnalysisAI ModelsAugust 22, 2026

GLM-5.3 matches Claude Fable 5 on DeepSWE accuracy at 5.4x lower cost

GLM-5.3 and Claude Fable 5 achieved nearly identical pass@1 scores of 69.0% and 69.7% on the DeepSWE benchmark. GLM-5.3 costs $3.99 per rollout compared to $21.63 for Fable 5, while outperforming it on pass@4 metrics with 87.6% versus 84.1%.

1 source

More stories today

Open the live feed