Claude 4 Opus Pricing 2026

Anthropic just slashed Claude 4 Opus prices by 50%. Here's what changed and whether you should switch.

April 22, 2026 — Anthropic announced a 50% price cut for Claude 4 Opus, bringing it in direct competition with GPT-4o, Qwen 3.6 Plus, and other top-tier models. This is the first major price adjustment since Claude 4's launch in January 2026.

New Claude 4 Opus Pricing (Effective April 2026)

Model	Input	Output	Context Window
Claude 4 Opus	$5.00/M	$15.00/M	200K
Claude 4 Opus (Before)	$10.00/M	$30.00/M	200K
Claude 4 Sonnet	$1.50/M	$7.50/M	200K
Claude 4 Haiku	$0.40/M	$2.00/M	200K

The 50% cut applies to both Claude 4 Opus and Claude 4 Sonnet (Sonnet dropped from $3/$15 to $1.50/$7.50). Haiku pricing remains unchanged.

How Claude 4 Opus Compares to Competitors

Model	Input	Output	Context	Best For
Claude 4 Opus	$5.00	$15.00	200K	Reasoning, coding
GPT-4o	$2.50	$10.00	128K	General purpose
Qwen 3.6 Plus	$0.33	$1.95	1M	Budget, long context
Gemini 2.0 Pro	$1.25	$5.00	2M	Massive context
DeepSeek-V3	$0.14	$0.28	64K	Cheap coding

Key takeaways:

• Claude 4 Opus is now 2x more expensive than GPT-4o on input tokens
• It costs 15x more than Qwen 3.6 Plus on input tokens
• But Claude 4 Opus still leads on coding benchmarks (AIME 2026: 92.3 vs GPT-4o's 89.1)

Should You Switch to Claude 4 Opus?

✅ Yes, if:

• You need best-in-class reasoning for complex coding or analysis tasks
• Your usage is moderate (under 1M tokens/month)
• You're currently using GPT-4o and want a quality upgrade without a huge price jump

❌ No, if:

• You're running high-volume agent workloads (use Qwen 3.6 Plus or DeepSeek-V3)
• You need massive context windows (Gemini 2.0 Pro has 2M vs Claude's 200K)
• You're on a tight budget (Qwen costs 93% less per token)

The Real Winner: Anthropic's Strategy

By cutting prices 50%, Anthropic is doing three things:

1. Undercutting OpenAI — GPT-5 is rumored to launch at $5-7/M input. Claude 4 Opus at $5 gets there first.
2. Pushing mid-tier models out — Sonnet at $1.50 makes older models (GPT-4, older Qwen) look overpriced.
3. Building enterprise lock-in — Lower entry costs mean more teams start with Claude, then scale up with Anthropic's ecosystem.

For most teams, the math is simple: if you need top-tier quality, Claude 4 Opus is now the best value at the high end. If you need volume, stay with Qwen or DeepSeek.

Related Tools

Compare AI model pricing in real-time with our AEO Checker — track which models mention your brand and at what cost.

Last updated: April 22, 2026 | Source: Anthropic official announcement, OpenRouter API