How much does DeepSeek v4 flash cost in comparison to a $200 codex GPT 5.5 High run?

I'm scared to just use an API and let it run rampant but I'm working on a few projects that are at that 80-90% mark where the last few % takes A LOT of work. I've gone through about 70% of my $200 Codex plan. I have GPT 5.5xhigh as the thinking model then have it going to 5.4-mini for 5.3-codex spark for the actual coding. WHich helps stretch things. But wondering if you dropped the same project /goal in into deepseek v4 flash using OpenCode how far would the same $200 get me? I know it's not exactly apples