Llm Gateway Cost And Caching easy
A feature sends 1,500 input tokens and receives 600 output tokens per request. Output tokens cost about 5x input tokens. Which single change cuts cost the most?
- A Trim the prompt from 1,500 to 1,200 input tokens by shortening the instructions.
- B Cap the response at 300 output tokens with a max-tokens limit and terse format.
- C Remove the few-shot examples so the prompt shrinks by about 400 input tokens.
- D Compress whitespace and punctuation in the prompt to save about 100 input tokens.
Answer: B. Cap the response at 300 output tokens with a max-tokens limit and terse format.
In input-token equivalents the request costs 1,500 + 5 x 600 = 4,500; cutting 300 output tokens saves 1,500 units, far more than the 300 or 400 saved by any prompt trim. **Output length is usually the main cost lever**, traded off against answer completeness. Trimming input wins only when prompts are huge and outputs are tiny, as in classification.
One system design question a day, with spaced repetition so the ones you miss come back.
Get the appRelated questions
- Your primary LLM provider is fully down for an hour, and no tested fallback exists for your agent feature. The product also has an FAQ answer feature backed by a semantic cache. What is the best degradation mode?
- Your gateway applies one 5-second timeout to all LLM calls. Classification calls work fine, but long streamed answers are frequently cut off. What is the best fix?
- A multi-tenant B2B support assistant answers questions using each customer's private knowledge base. To raise the semantic cache hit rate from 22% to 35%, an engineer proposes one shared cache across all tenants. What is the main problem?
- During a provider brownout returning 429 and 503 errors, the client SDK retries 3 times and the gateway also retries 3 times, both immediately. Provider load spikes, and the bill grows from partially streamed responses. What should you change?
- Finance asks for LLM spend broken down by tenant and by product feature. Today the gateway only sees the total on the provider's monthly invoice. What should you do?
- A retail assistant uses a semantic cache. Weeks after a store changed its holiday hours, users still receive the old hours from cached answers. What is the right fix?