Processing 25,000 lines of code (~280,000 tokens) with Qwen3.8 27B (OpenRouter) costs $0.1176 for codebase ingestion and $0.2856 for an AI-powered code review and refactoring pass.
Quick answer
A 25,000 lines of code codebase is estimated at 280,000 tokens for Qwen3.8 27B (OpenRouter). Ingestion costs $0.1176; a review and refactoring pass with roughly 20% output costs about $0.2856.
Method & trust
Code is estimated at a tokenizer-specific tokens-per-line ratio, then priced separately for repository context and generated review output. Comments, minified files, tests, and repeated context can materially change the bill.
Feed 25,000 lines of code into the prompt context for repository search, Q&A, or architecture planning.
Ingest 25,000 lines of code and generate audit findings, unit test recommendations, and refactor diffs.
Qwen3.8 27B (OpenRouter) writes 25,000 lines of code from scratch based on product specifications.
| Model | Provider | Code Ingestion | Cached Ingestion | Code Review Cost | Context Limit |
|---|---|---|---|---|---|
| Qwen3.8 27B (OpenRouter) (Current) | qwen | $0.1176 | $0.0294 | $0.2856 | 1,000,000 |
| GPT-5.6 Terra | openai | $0.51 | $0.051 | $1.122 | 1,050,000 |
| GPT-5.4 Workhorse | openai | $0.6375 | $0.0638 | $1.402 | 256,000 |
| o3-mini | openai | $0.2805 | $0.1403 | $0.5049 | 200,000 |
| o4-mini | openai | $0.2805 | $0.0701 | $0.5049 | 256,000 |
| o1-mini | openai | $0.2805 | $0.1403 | $0.5049 | 128,000 |
On average, code yields approximately 11.2 tokens per line in Qwen3.8 27B (OpenRouter) (qwen_bpe tokenizer). Indentation, brackets, camelCase variable names, and comments slightly increase token density compared to plain English text. 25,000 lines of code produces approximately 280,000 tokens.
Sending 25,000 lines of code as context and generating a thorough code review with recommendations costs approximately $0.2856. Utilizing prompt caching on repeat turns or static repository definitions drops this to $0.1974.
Qwen3.8 27B (OpenRouter) has a context window of 1,000,000 tokens. 25,000 lines of code consumes 28.00% of its total available context.