cost-optimization
# Claude Code Cost Optimization
What this skill does
# Claude Code Cost Optimization
Complete guide to managing costs, model routing, token usage, and caching.
## Cost Tracking
### /cost Command
```
/cost
```
Shows:
- Input tokens consumed
- Output tokens consumed
- Cache read tokens (cheaper)
- Cache write tokens
- Total estimated cost (USD)
## Model Selection & Routing
### Available Models
| Model | ID | Best For | Cost |
|-------|-----|---------|------|
| Opus 4.6 | `claude-opus-4-6` | Architecture, complex decisions | Highest |
| Sonnet 4.6 | `claude-sonnet-4-6` | General development, implementation | Medium |
| Haiku 4.5 | `claude-haiku-4-5-20251001` | Quick lookups, simple tasks | Lowest |
### Switching Models
```
/model claude-haiku-4-5-20251001 # Switch to Haiku for simple tasks
/model claude-sonnet-4-6 # Switch back to Sonnet
/model claude-opus-4-6 # Switch to Opus for complex work
```
### CLI Model Override
```bash
claude -m claude-haiku-4-5-20251001 -p "quick question"
```
### Settings Configuration
```json
{
"model": "claude-sonnet-4-6",
"smallFastModel": "claude-haiku-4-5-20251001"
}
```
## Token Reduction Strategies
### 1. Use /compact Frequently
```
/compact # Compress full conversation
/compact focus on the API # Compress with specific focus
```
Reduces context window size, lowering per-message input costs.
### 2. Targeted File Reads
```
// Expensive: read entire large file
Read(file_path="large-file.ts") // ~5000 tokens
// Cheap: read specific section
Read(file_path="large-file.ts", offset=100, limit=30) // ~300 tokens
// Cheap: search first
Grep(pattern="function auth", path="src/") // ~100 tokens
```
### 3. Use Sub-Agents for Research
Sub-agents process information internally and return summaries:
```
// Main context gets only the summary (~500 tokens)
// Instead of 20 file reads (~50,000 tokens)
Agent(subagent_type="Explore", prompt="Find all database models")
```
### 4. Grep Before Read
```
// Don't read every file looking for something
// Search first, then read only matching files
Grep(pattern="TODO|FIXME", type="ts")
```
### 5. Background Tasks
```
// Long tasks don't consume main context while running
Agent(run_in_background=true, ...)
Bash(command="npm test", run_in_background=true)
```
### 6. Clear Between Unrelated Tasks
```
/clear # Reset context for new topic
```
## Prompt Caching
### How It Works
- Claude Code caches system prompts and conversation history
- Cached tokens cost significantly less (~90% savings)
- Cache hits happen when the same prefix appears in consecutive requests
### Maximizing Cache Hits
1. **Keep CLAUDE.md stable** — Changes invalidate cache
2. **Consistent system prompts** — Don't modify with `--append-system-prompt` frequently
3. **Sequential conversations** — Cache benefits multi-turn conversations
### API-Level Caching
```typescript
const response = await client.messages.create({
model: "claude-sonnet-4-6",
max_tokens: 1024,
system: [
{
type: "text",
text: "Your system prompt here...",
cache_control: { type: "ephemeral" }
}
],
messages: [...]
});
// Usage shows cache info
console.log(response.usage.cache_creation_input_tokens);
console.log(response.usage.cache_read_input_tokens);
```
## Provider Cost Comparison
### Anthropic Direct
Standard pricing, most features.
### AWS Bedrock
```bash
CLAUDE_CODE_USE_BEDROCK=1 claude
```
- May have different pricing through AWS agreements
- Cross-region inference available
- Committed use discounts possible
### Google Vertex AI
```bash
CLAUDE_CODE_USE_VERTEX=1 claude
```
- GCP pricing and billing
- May have committed use discounts
## Batch Processing (50% Savings)
For non-interactive workloads, use the Message Batches API:
```typescript
const batch = await client.messages.batches.create({
requests: [
{
custom_id: "review-1",
params: {
model: "claude-sonnet-4-6",
max_tokens: 1024,
messages: [{ role: "user", content: "Review file1.ts" }]
}
},
{
custom_id: "review-2",
params: {
model: "claude-sonnet-4-6",
max_tokens: 1024,
messages: [{ role: "user", content: "Review file2.ts" }]
}
}
]
});
```
Batch processing gives 50% cost reduction with 24-hour SLA.
## Cost Estimation
### Rule of Thumb
| Task | Approximate Cost |
|------|-----------------|
| Simple question | $0.01 - $0.05 |
| Code review (1 file) | $0.05 - $0.15 |
| Feature implementation | $0.20 - $1.00 |
| Complex refactoring | $0.50 - $2.00 |
| Full project analysis | $1.00 - $5.00 |
### Factors Affecting Cost
1. Model used (Opus > Sonnet > Haiku)
2. Context window size (more history = more input tokens)
3. Number of tool calls (each adds output + input)
4. File sizes read
5. Number of conversation turns
6. Extended thinking budget
## Best Practices
1. **Start with Sonnet** — Good balance of quality and cost
2. **Use Haiku for exploration** — Switch for simple lookups
3. **Upgrade to Opus for architecture** — Worth the cost for complex decisions
4. **Compact regularly** — Smaller context = lower cost per turn
5. **Delegate research** — Sub-agents are cost-neutral for main context
6. **Monitor with /cost** — Track spending periodically
7. **Use batch API for bulk** — 50% savings on non-interactive work
8. **Leverage caching** — Keep system prompts stable
Related in AI Agents
skill-development
IncludedComprehensive meta-skill for creating, managing, validating, auditing, and distributing Claude Code skills and slash commands (unified in v2.1.3+). Provides skill templates, creation workflows, validation patterns, audit checklists, naming conventions, YAML frontmatter guidance, progressive disclosure examples, and best practices lookup. Use when creating new skills, validating existing skills, auditing skill quality, understanding skill architecture, needing skill templates, learning about YAML frontmatter requirements, progressive disclosure patterns, tool restrictions (allowed-tools), skill composition, skill naming conventions, troubleshooting skill activation issues, creating custom slash commands, configuring command frontmatter, using command arguments ($ARGUMENTS, $1, $2), bash execution in commands, file references in commands, command namespacing, plugin commands, MCP slash commands, Skill tool configuration, or deciding between skills vs slash commands. Delegates to docs-management skill for official documentation.
reprompter
IncludedTransform messy prompts into well-structured, effective prompts — single or multi-agent. Use when: "reprompt", "reprompt this", "clean up this prompt", "structure my prompt", rough text needing XML tags and best practices, "reprompter teams", "repromptception", "run with quality", "smart run", "smart agents", multi-agent tasks, audits, parallel work, anything going to agent teams. Don't use when: simple Q&A, pure chat, immediate execution-only tasks. See "Don't Use When" section for details. Outputs: Structured XML/Markdown prompt, quality score (before/after), optional team brief + per-agent sub-prompts, agent team output files. Success criteria: Single mode quality score ≥ 7/10; Repromptception per-agent prompt quality score 8+/10; all required sections present, actionable and specific.
adaptive-compaction
IncludedAdaptive add-on policy and recovery layer that decides WHEN to compact, prune, snapshot, or fork -- replacing fixed-percent auto-compaction across Claude Code, Codex, and MCP-capable hosts. Trigger on auto-compact timing or damage: "when should I compact", "is it safe to compact now or start a fresh session", "auto-compact fires too early/mid-task", "switching to an unrelated task but the window still has space", "context rot", "answers get worse the longer the session runs", "the agent forgot the plan or my decisions after it summarized", "add a layer on top that manages context without changing the agent", raising autoCompactWindow to give the policy room, or installing/tuning a cross-tool compaction policy or PreCompact hook -- even when "compaction" is never said but the problem is context-window pressure or post-summarization memory loss. Do NOT use to summarize a conversation, build RAG, write a summarization prompt (decides WHEN not HOW), or answer max-context-length trivia.
agent-skill-creator
IncludedCreate cross-platform agent skills from workflow descriptions. Activates when users ask to create an agent, automate a repetitive workflow, create a custom skill, or need advanced agent creation. Triggers on phrases like create agent for, automate workflow, create skill for, every day I have to, daily I need to, turn process into agent, need to automate, create a cross-platform skill, validate this skill, export this skill, migrate this skill. Supports single skills, multi-agent suites, transcript processing, template-based creation, interactive configuration, cross-platform export, and spec validation.
llm-wiki
IncludedUse when building or maintaining a persistent personal knowledge base (second brain) in Obsidian where an LLM incrementally ingests sources, updates entity/concept pages, maintains cross-references, and keeps a synthesis current. Triggers include "second brain", "Obsidian wiki", "personal knowledge management", "ingest this paper/article/book", "build a research wiki", "compound knowledge", "Memex", or whenever the user wants knowledge to accumulate across sessions instead of being re-derived by RAG on every query.
skill-master
IncludedAgent Skills authoring, evaluation, and optimization. Create, edit, validate, benchmark, and improve skills following the agentskills.io specification. Use when designing SKILL.md files, structuring skill folders (references, scripts, assets), ingesting external documentation into skills, running trigger evals, benchmarking skill quality, optimizing descriptions, or performing blind A/B comparisons. Keywords: agentskills.io, SKILL.md, skill authoring, eval, benchmark, trigger optimization.