Unofficial guide. Not affiliated with Meta.

Last updated:

Claude Opus 5.5: price, free access and real cost per task

Anthropic released Claude Opus 5.5 on September 22, 2026. It is cheaper than Opus 5 and, on Anthropic's own table, beats Fable 5.1 on coding. Here is what it costs to actually use.

Verdict Opus 5.5 is the model for hard, long-running coding work. At $4.00 / $20.00 per 1M tokens it is 20% cheaper per token than Opus 5 and far cheaper than Fable 5.1 ($10.00 / $50.00). It is not on the free Claude plan. For everyday tasks, Sonnet 5.5 costs half as much and scored higher on Terminal-Bench 4.0.

Key facts

API model IDclaude-opus-5-5
ReleasedSeptember 22, 2026
Context window1M tokens
Max output128K tokens (300K on the Batch API, beta)
InputText and images (no video)
Available onClaude API, AWS, Google Cloud, Microsoft Azure, Claude apps (paid plans) and Claude Code

Source: Anthropic: Opus 5.5 model page

Claude Opus 5.5 API pricing

USD per 1M tokens, from Anthropic's model page and launch post.

Price per 1M tokensOpus 5.5Opus 5
Input$4.00$5.00
Output$20.00$25.00
Cache write (5 min)$5.00$6.25
Cache hit$0.20$0.50

Source: Anthropic: Introducing Claude Opus 5.5

Is Claude Opus 5.5 free?

Source: Claude plans and pricing

Using Opus 5.5 in Claude Code

What one coding task costs on Opus 5.5

Same assumptions as our calculator, list prices, per agent task:

Codebase sizeOpus 5.5Sonnet 5.5Fable 5.1
Small (< 10k lines)$0.223$0.115$0.549
Medium (10k-100k lines)$0.498$0.258$1.222
Large / monorepo (> 100k lines)$1.188$0.618$2.910

At 173 tasks a month (medium codebase) that is about $86 a month on Opus 5.5, versus $45 on Sonnet 5.5.

Compare every model in the cost calculator Opus 5.5 vs Sonnet 5.5: which one to use

Benchmarks (Anthropic's numbers)

From Anthropic's launch post. Opus 5.5 at max effort (Terminal-Bench at xhigh).

BenchmarkOpus 5.5Fable 5.1Opus 5GPT-6 Astra
Terminal-Bench 4.066.4%55.8%52.3%57.9%
FrontierCode 1.154.4%50.3%48.0%53.3%
CursorBench 4.057.8%51.8%46.6%—
AutomationBench40.0%31.4%26.9%41.4%
Humanity's Last Exam67.7%65.6%63.6%57.2%
OSWorld 2.181.8%80.7%74.0%—

For comparison, Anthropic reports Sonnet 5.5 at 70.6% on Terminal-Bench 4.0.

Source: Anthropic: Introducing Claude Opus 5.5

FAQ

How much does Claude Opus 5.5 cost?

$4.00 per 1M input tokens and $20.00 per 1M output tokens on the API, with cache hits at $0.20. Opus 5 was $5.00 / $25.00. Anthropic says it costs about 40% less to run than Opus 5 on typical workloads.

Is Claude Opus 5.5 free?

No. The free Claude plan includes Sonnet and Haiku, not Opus. You need Pro or above, or an API key.

Can Opus 5.5 read or make videos?

No. Anthropic lists text and images as input and text as output. It cannot take video input or generate video.

Is there a Claude Fable 5.5?

No. The newest Fable model is Fable 5.1 ($10.00 / $50.00 per 1M tokens). Anthropic says Opus 5.5 performs at Fable 5.1's level on most work at a lower price.

Is Claude Haiku 5.5 out?

Not yet. Anthropic said on 2026-09-22 that Haiku 5.5 will follow "in the coming weeks". There is no price or model ID yet.

Independent page. Not affiliated with Anthropic. Benchmark figures are Anthropic's; we have not run our own tests.