Open Calculator →

Claude 4 Pricing — All Models

Opus 4.7 · Sonnet 4.6 · Haiku 4.5 — full API rates including cache pricing, practical cost examples, and a free log analyzer.

Calculate My Session Cost →

Claude 4 Model Pricing at a Glance

Model Input / 1M tokens Output / 1M tokens Cache Write Cache Read Best for
Claude Opus 4.7 Most Capable $15.00 $75.00 $18.75 $1.50 Complex reasoning, research
Claude Sonnet 4.6 Default $3.00 $15.00 $3.75 $0.30 Claude Code, coding tasks
Claude Haiku 4.5 Cheapest $0.80 $4.00 $1.00 $0.08 Fast subtasks, high volume

Detailed Pricing by Model

Claude Opus 4.7 claude-opus-4-7
Input: $15.00 / 1M tokens
Output: $75.00 / 1M tokens

Cache write: $18.75 / 1M
Cache read: $1.50 / 1M 90% off input

Best for hard reasoning tasks, complex code architecture, research synthesis. Use sparingly — at 5× Sonnet's price, reserve it for tasks where quality truly matters.

Claude Haiku 4.5 claude-haiku-4-5
Input: $0.80 / 1M tokens
Output: $4.00 / 1M tokens

Cache write: $1.00 / 1M
Cache read: $0.08 / 1M 90% off input

Ideal for fast sub-tasks, classification, data extraction, and any high-volume workload where you need speed at minimal cost. Nearly 4× cheaper than Sonnet per token.

Claude 4 vs Claude 3 Pricing

Model Generation Input / 1M Output / 1M Cache Read
Claude Opus 4.7 Claude 4 $15.00 $75.00 $1.50
Claude Opus 3 Claude 3 $15.00 $75.00 $1.50
Claude Sonnet 4.6 Claude 4 $3.00 $15.00 $0.30
Claude Sonnet 3.5 Claude 3 $3.00 $15.00 $0.30
Claude Haiku 4.5 Claude 4 $0.80 $4.00 $0.08
Claude Haiku 3 Claude 3 $0.25 $1.25 $0.03
Key takeaway: Claude 4 Opus and Sonnet match their Claude 3 equivalents in price while delivering significantly stronger capability — especially for agentic coding tasks. Haiku 4.5 is priced higher than Haiku 3 but offers substantially better reasoning and instruction-following. For most Claude Code users on Sonnet, migrating from 3.5 to 4.6 is a free upgrade.

Practical Claude 4 Cost Examples

One hour of Claude Code (Sonnet 4.6, typical)
~80K input + 20K output (30% cached) → input $0.17 + cached reads $0.02 + output $0.30 = ~$0.49
Full workday of heavy Claude Code (Sonnet 4.6, 8h)
~500K input + 150K output (60% cached) → $0.60 input + $0.09 cached + $2.25 output = ~$2.94
Occasional Opus 4.7 call for a hard problem (5K tokens total)
3K input + 2K output → $0.045 + $0.15 = ~$0.20 per hard call
10,000 document classifications (Haiku 4.5)
10K calls × 500 input + 50 output avg → 5M input + 0.5M output = $4.00 + $2.00 = $6.00 total
1,000 code reviews (Sonnet 4.6, 4K in / 1K out each)
4M input + 1M output → $12.00 + $15.00 = $27.00 total (or ~$12 with 70% cache hit)

Frequently Asked Questions

How much does Claude 4 cost?

Claude 4 pricing depends on the model tier: Opus 4.7 costs $15/M input and $75/M output. Sonnet 4.6 costs $3/M input and $15/M output. Haiku 4.5 costs $0.80/M input and $4/M output. All models support prompt caching, which can cut costs by up to 90% on repeated context.

Is Claude 4 cheaper than GPT-4o?

Claude Sonnet 4.6 and GPT-4o are comparably priced at ~$3/M input and $15/M output. Claude Haiku 4.5 ($0.80/$4) is significantly cheaper than GPT-4o mini ($0.15/$0.60 — though much less capable). Where Claude has a clear advantage: cache read prices are $0.30/M for Sonnet vs $1.25/M for GPT-4o cached input — a 4× cost difference on repeated prompts, which matters greatly for Claude Code sessions.

What is Claude 4 prompt caching and how much does it save?

Prompt caching lets you store context (system prompts, documents, prior conversation) and serve it at ~10% of the normal input token price. Cache writes are charged at ~1.25× input, then every cache hit is billed at just 10% of standard input price — a 90% saving. For Claude Code sessions with long system prompts, caching typically cuts the total input bill by 40–70%.

Which Claude 4 model should I use?

For Claude Code: Sonnet 4.6 (the default) handles nearly all coding tasks at $3/$15/M. Upgrade to Opus 4.7 only for complex architecture decisions or reasoning-heavy tasks. Use Haiku 4.5 for background tasks like file indexing or fast classifiers where you need sub-second latency.

For API workloads: start with Sonnet 4.6 unless you have budget constraints (use Haiku) or need maximum quality (use Opus).

How do I calculate my exact Claude 4 API costs?

If you use Claude Code, your session logs are at ~/.claude/projects/<project>/*.jsonl. Each line records the model, token counts (input, output, cache write, cache read), and timestamp. Paste the file into the Claude Code Cost Calculator — it applies the correct Claude 4 per-model pricing and shows a breakdown by model, tool call, and hour. For direct API usage, the same logs and pricing apply.

Are Claude 4 prices the same as Claude 3.5 prices?

Claude Sonnet 4.6 = Claude 3.5 Sonnet = $3/$15/M. Claude Haiku 4.5 is more expensive than Haiku 3 ($0.25/$1.25/M) but cheaper than Sonnet. Opus 4.7 = Opus 3 = $15/$75/M. Bottom line: upgrading from Sonnet 3.5 to Sonnet 4.6 is free — you get better performance at the same price.