Blog
LLM Cost Optimization Insights
Research, techniques, and production results for engineers reducing AI inference costs.
Best Local LLM for Coding (2026): 5 Models, Measured
→
How Much VRAM Do You Need for a Local LLM? Measured
→
Prompt Caching in vLLM and Ollama: Measured Speed and Energy
→
What Does One LLM Request Cost? Dollars, Tokens and Joules
→
SLM vs LLM: When Is a Small Language Model Good Enough?
→
Qwen 3.8 27B for Coding: VRAM, Speed and Pass Rate on R9700
→
Does LLM Self-Correction Work? What 7B Code Repair Shows
→
Qwen2.5-Coder-7B Hits 87.8% HumanEval+ on One 32GB GPU
→
Reasoning Token Costs: What You're Actually Paying For
→
Self-Hosting an LLM: Is It Actually Cheaper Than the API?
→
Best LLM Cost Tracking Tools (2026): A Vendor-Neutral Comparison
→
LLM Caching Strategies: Prompt, Semantic, and KV Cache — Which to Deploy First
→
The LLM Effective Cost Table: What Cache + Batch Pricing Actually Costs in 2026
→
Prompt Caching in 2026: OpenAI vs Claude vs Gemini Pricing
→
LLM Batch API: The Flat 50% Discount and How It Stacks With Caching
→
Semantic Caching for LLMs: Real Hit Rates, Thresholds, and Limits
→
Long-Context LLM Cost Management: Stop Paying for Stale Tokens
→
AI Agent Cost Optimization: 7 Patterns That Blow Up Your Budget
→
Featured
LLM Cost Optimization: Why Enterprises Overspend 50–90% and How to Fix It
→
LLM Model Routing: Automatically Send Every Query to the Cheapest Capable Model
→
Prompt Compression: Reduce LLM Token Costs by 2–20× Without Losing Quality
→
LoRA Adapters: Fine-Tune LLMs for 1% of the Cost of Full Fine-Tuning
→
Follow the measurements
New measurements when we publish them, roughly monthly. No sales call.