Optimization Articles
Browse 263 articles about Optimization.
Mixture of Experts Architecture Explained: How GLM 5.2 Runs 40B Active Parameters
GLM 5.2 has 744B total parameters but only 40B active per token thanks to MoE routing. Learn how this architecture enables local inference on consumer hardware.
How to Run a 744B AI Model on a Consumer Laptop Using Colibri
Colibri uses three-tier memory and SSD streaming to run GLM 5.2 on consumer hardware. Learn how the hot-cold expert split makes this possible.
What Is 1-Bit Quantization for AI Models? How Cactus Bonsai Runs 27B Parameters on a Phone
Cactus Bonsai compresses a 27B parameter model to 3.9GB using 1-bit quantization and quantization-aware training. Learn how it works and what it enables.
AI Agent Harness Bloat: How to Audit and Clean Your Claude or ChatGPT Setup
Discover how accumulated rules, skills, and instructions degrade AI agent performance—and learn a 6-principle framework to clean your harness before it breaks.
How to Use GPT-5.6 for Agentic Coding: Real-World Results and Cost Comparison
GPT-5.6 Soul delivers near-Fable-5 quality at a fraction of the cost. See real benchmarks, cost-per-task comparisons, and when to choose it over Claude.
How to Use the Advisor-Executor Pattern: Plan with Fable 5, Build with Sonnet
Cut AI costs by 50% using the advisor-executor pattern. Use Fable 5 for planning and code review, then switch to Sonnet for implementation and execution.
Local AI vs Cloud AI: Open-Weight Models, Licensing, and the Hybrid Routing Strategy
71% of ChatGPT queries could run locally, but open-weight licensing is a minefield. Learn the three tiers of local AI and when hybrid routing saves money.
Local AI vs Cloud AI for Agents: The Hybrid Routing Strategy That Saves Money
Learn when to run AI models locally and when to route to cloud APIs. A practical hybrid strategy for reducing costs without sacrificing quality.
How to Use GPT-5.6 Sol as an Orchestrator with Cheaper Sub-Agent Models
GPT-5.6 Sol excels at planning and orchestration. Learn how to pair it with Luna or Terra sub-agents to cut costs while keeping output quality high.
Plan with Fable 5, Build with Sonnet: The Model Routing Pattern That Cuts AI Costs
Use Claude Fable 5 for planning and architecture, then switch to Sonnet for implementation. This routing pattern cuts token costs without sacrificing quality.
How to Reduce Claude Fable 5 Token Costs: 8 Settings to Change Right Now
Fable 5 now bills at API rates. Learn which settings to disable—MCP servers, auto-memory, Claude.md bloat—to cut token costs without losing quality.
How to Use Progressive Disclosure in AI Agent Design to Scale Capabilities
Progressive disclosure lets agents load only the instructions they need per query. Learn how Pydantic AI 2.0 implements this pattern to prevent context bloat.
Token Efficiency vs Raw Intelligence: Why GPT-5.6 Beats Claude Fable 5 on Cost-Per-Result
GPT-5.6 Sol uses half the tokens of Claude Fable 5 and costs 3x less. Here's why token efficiency often matters more than benchmark scores for real workflows.
How to Use the Advisor-Executor Pattern in Claude Code to Extend Your Fable 5 Limit
Use Fable 5 as an advisor and Opus or Sonnet as executor to get 93% of the work done at a fraction of the token cost—with a real example.
Claude Code Effort Levels: Why Max Mode Often Hurts More Than It Helps
Independent testing shows effort levels below max barely change output quality. Learn when to use default, high, and max effort in Claude Code.
How to Use Claude Code's /fewer Permission Prompt to Build a Custom Allow List
The /fewer permission prompt scans your session history to auto-generate a tailored allow list—giving you the speed of auto-mode with more control.
What Is Context Rot in AI Agents and How Does Auto-Compact Fix It?
Context rot degrades AI agent quality at 70-80% context fill. Learn how to set Claude Code's auto-compact threshold to prevent it before it starts.
What Is the 'Fable Mode' Skill? How to Make Cheaper Models Think Like Frontier AI
The Fable Mode skill extracts Claude Fable 5's reasoning habits and injects them into cheaper models like Opus. Here's how it works and how to build it.
How to Use Model Routing to Cut AI Agent Costs by 60%
Learn how to route tasks to cheaper models like Sonnet and Haiku instead of always using frontier models, without sacrificing output quality.
AI Model Routing: How to Cut Costs 60% by Matching Tasks to the Right Model
Learn how to route AI tasks to the right model tier—frontier for planning, cheaper for execution—and cut your AI bill by up to 60%.