Correctly calculate remaining context size (#3190)

We had multiple issues with context size calculation:
1. `initial_prompt_tokens` calculation based on cache size is not
reliable, cache misses might set it to much higher value. For now
hardcoded to a safer constant.
2. Input context size for GPT-5 is 272k (that's where 33% came from).

Fixes.

codex-rs/core/src/openai_model_info.rs

@@ -79,12 +79,12 @@ pub(crate) fn get_model_info(model_family: &ModelFamily) -> Option<ModelInfo> {
         }),
 
         "gpt-5" => Some(ModelInfo {
-            context_window: 400_000,
+            context_window: 272_000,
             max_output_tokens: 128_000,
         }),
 
         _ if slug.starts_with("codex-") => Some(ModelInfo {
-            context_window: 400_000,
+            context_window: 272_000,
             max_output_tokens: 128_000,
         }),