Claude Pricing in 2026: API Costs for Opus, Sonnet, and Haiku
Last updated: April 9, 2026 · Official vendor pricing verified against Anthropic pricing documentation. EvoLink route pricing reflects current site-supported channel pricing and may differ from direct Anthropic billing.
TL;DR
Claude API pricing depends on which model you pick and which pricing scope you mean. This guide separates:
- official Anthropic API pricing
- long-context / caching / batch pricing rules
- current EvoLink route pricing, where available
Here are the current official base rates per million tokens (MTok):
| Model | Input | Output | Best for |
|---|---|---|---|
| Claude Opus 4.6 | $5.00 | $25.00 | Flagship reasoning, complex coding, agents |
| Claude Opus 4.5 | $5.00 | $25.00 | Previous-gen flagship, stable versioned ID |
| Claude Opus 4.1 | $15.00 | $75.00 | Agentic coding, multi-file refactors |
| Claude Sonnet 4.6 | $3.00 | $15.00 | Best balance of speed/cost/intelligence |
| Claude Sonnet 4.5 | $3.00 | $15.00 | Recommended default, strong coding |
| Claude Sonnet 4.0 | $3.00 | $15.00 | Production workloads, 1M context beta |
| Claude Haiku 4.5 | $1.00 | $5.00 | Fastest, cheapest — high-volume workloads |
If you are comparing direct vendor pricing with gateway pricing, Anthropic's official rates are the baseline. EvoLink can expose lower route-specific pricing for the same Claude family, but that is a separate pricing scope from direct Anthropic billing.
Claude Opus 4.6 Pricing
effort parameter for cost/quality control.| Component | Rate (per MTok) |
|---|---|
| Input tokens | $5.00 |
| Output tokens | $25.00 |
| Cache write (5 min) | $6.25 |
| Cache write (1 hour) | $10.00 |
| Cache read | $0.50 (0.1x input) |
| Batch input | $2.50 (50% off) |
| Batch output | $12.50 (50% off) |
claude-opus-4-6When Opus 4.6 is worth the premium
- Complex multi-step coding (80.8% SWE-bench Verified)
- Agent orchestration with tool use
- Research synthesis requiring extended thinking
- Tasks where quality matters more than speed
Claude Opus 4.5 Pricing
| Component | Rate (per MTok) |
|---|---|
| Input tokens | $5.00 |
| Output tokens | $25.00 |
| Cache write (5 min) | $6.25 |
| Cache read | $0.50 |
claude-opus-4-5-20250514Claude Opus 4.1 Pricing
| Component | Rate (per MTok) |
|---|---|
| Input tokens | $15.00 |
| Output tokens | $75.00 |
| Cache write (5 min) | $18.75 |
| Cache read | $1.50 |
claude-opus-4-1-20250805Is Opus 4.1 still worth using?
Opus 4.1 costs 3x more than Opus 4.6 per token. Use it only if your workflow specifically benefits from its older route characteristics or you need to keep a pinned version for compatibility. For most new workloads, Opus 4.6 at $5/$25 is the stronger value route.
Claude Sonnet 4.6 Pricing
| Component | Rate (per MTok) |
|---|---|
| Input tokens | $3.00 |
| Output tokens | $15.00 |
| Cache write (5 min) | $3.75 |
| Cache read | $0.30 |
| Batch input | $1.50 (50% off) |
| Batch output | $7.50 (50% off) |
claude-sonnet-4-6Claude Sonnet 4.5 Pricing
| Component | Rate (per MTok) |
|---|---|
| Input tokens | $3.00 |
| Output tokens | $15.00 |
| Cache write (5 min) | $3.75 |
| Cache read | $0.30 |
claude-sonnet-4-5-20250929Claude Sonnet 4.0 Pricing
| Component | Rate (per MTok) |
|---|---|
| Input tokens | $3.00 |
| Output tokens | $15.00 |
| Input (>200K, 1M beta) | $6.00 |
| Output (>200K, 1M beta) | $22.50 |
| Cache write (5 min) | $3.75 |
| Cache read | $0.30 |
claude-sonnet-4-0-20250514Claude Haiku 4.5 Pricing
| Component | Rate (per MTok) |
|---|---|
| Input tokens | $1.00 |
| Output tokens | $5.00 |
| Cache write (5 min) | $1.25 |
| Cache read | $0.10 |
| Batch input | $0.50 (50% off) |
| Batch output | $2.50 (50% off) |
claude-haiku-4-5-20251001When to choose Haiku over Sonnet
- Free-tier products where cost per request matters
- Parallelized sub-agents and high-volume pipelines
- Latency-sensitive chat and support agents
- Tasks where Sonnet 4 quality is sufficient (73.3% SWE-bench)
Cost Comparison: Claude vs GPT vs Gemini
How does Claude pricing compare to other frontier models? Here's a side-by-side:
| Model | Input/MTok | Output/MTok | Context | Max Output |
|---|---|---|---|---|
| Claude Opus 4.6 | $5.00 | $25.00 | 1M | 64K |
| Claude Sonnet 4.6 | $3.00 | $15.00 | 1M | 128K |
| Claude Haiku 4.5 | $1.00 | $5.00 | 200K | — |
| GPT-5.4 | $2.50 | $20.00 | 1M | 128K |
| GPT-5.2 | $1.75 | $14.00 | 400K | 128K |
| Gemini 3.1 Pro | $2.00 | $12.00 | 1M | 64K |
How to Reduce Claude API Costs
1. Use prompt caching
If you reuse system prompts or long context across requests, caching can cut input costs by up to 90%. Cache reads are billed at 0.1x the base input rate.
2. Use batch processing
Batch API provides a 50% discount on both input and output tokens for asynchronous jobs. Ideal for nightly automation, data processing, and non-real-time workloads.
3. Pick the right model tier
Don't use Opus for tasks Sonnet can handle. Don't use Sonnet for tasks Haiku can handle. Match model capability to task complexity.
4. Compare official pricing with EvoLink route pricing
- Claude Opus 4.6: from $4.50/MTok input (−10%)
- Claude Sonnet 4.6: from $2.70/MTok input (−10%)
- Claude Haiku 4.5: from $0.90/MTok input (−10%)
Use direct Anthropic pricing when budgeting for vendor-native usage. Use EvoLink route pricing when you specifically plan to buy Claude through EvoLink's OpenAI-compatible endpoint.
5. Monitor and set budgets
Use Anthropic's usage dashboard or EvoLink's built-in monitoring to track spend by model, set alerts, and avoid surprise bills.
FAQ
How much does Claude API cost?
Claude API pricing varies by model. On Anthropic's official API pricing, Haiku 4.5 starts at $1/$5 per million tokens, Sonnet 4.6 is $3/$15, and Opus 4.6 is $5/$25. Prompt caching and batch processing can reduce effective costs further. All figures here are per million tokens as of April 9, 2026.
What is the cheapest Claude API model?
Claude Haiku 4.5 at $1 input / $5 output per million tokens. It is the lowest-cost current Claude model for high-volume production workloads.
Is Claude API free?
There is no permanent free Claude API tier for production usage. Anthropic's API pricing is usage-based, and any trial or promotional credit availability should be checked against Anthropic's current billing pages. EvoLink also advertises free trial credits for new accounts on current site messaging.
How does Claude pricing compare to GPT?
Claude Sonnet 4.6 ($3/$15) sits above GPT-5.2 ($1.75/$14) and Gemini 3.1 Pro ($2/$12) on direct vendor pricing, while Claude Opus 4.6 ($5/$25) is the premium Claude tier. Through EvoLink, route pricing can narrow that gap, but those prices should be treated separately from official Anthropic billing.
What is prompt caching and how much does it save?
Prompt caching lets you reuse frequently sent context (system prompts, documents) across requests. Cache writes cost 1.25x the base input rate, but cache reads cost only 0.1x — saving up to 90% on repeated input tokens. This is especially valuable for agents and chatbots with long system prompts.
Does Claude have batch processing discounts?
Yes. Anthropic's Batch API provides a 50% discount on both input and output tokens. Jobs are processed asynchronously within 24 hours. This is ideal for data pipelines, content generation, and any workload that doesn't need real-time responses.
Where can I access Claude API at the best price?
Start Using Claude API on EvoLink
Ready to integrate Claude into your product? EvoLink gives you access to every Claude model through a single OpenAI-compatible endpoint — with discounted route pricing, built-in rate limit handling, and usage monitoring.
- Browse all Claude models →
- Get your API key → (evolink.ai)
- Claude Opus 4.6 → | Claude Sonnet 4.6 → | Claude Haiku 4.5 →
