New GPT models

Prompt

I want to compare these models in agentic coding. Include cost per task too. Claude: Opus 4.8 Opus 4.5 Sonnet 4.5 Haiku 4.5 GPT: 6.1 Sol 6 Astra 6 Luna 5.6 Sol 5.6 Luna Highlight these: 6.1 Sol 6 Astra 6 Luna cause they are the newest.