GLM-5.3 Flash: 17x cheaper, 5.6 points behind on DeepSWE
Together AI ran 900 DeepSWE rollouts: GLM-5.3 Flash trails the full model by 5.6 points at pass@1 but only 2.6 at pass@4, at 17x lower cost ($0.24 vs $3.99 per rollout). A cascade routing solves 80.9% of tasks at $1.70 each.
How this story unfolded
3 days · 4 reports · 6 community posts · from Aug 29
- Aug 29
- Aug 30
- Sep 1
AI Models by email
Get an email when there's news on AI Models
No news that day, no email.
More stories today
- Security researcher changes mind on AI guardrails
- Forescout uses Claude AI to port PLC exploit in hours
- Weaviate shows how to extract meaning from charts and tables in PDFs
- Alok launches AI-powered personalized music video campaign for WAAW headphones
- X launches MCP server for advertiser tools