By the toolist.ai editors. Researched from public sources, not hands-on tested by toolist.ai. Updated November 3, 2026; prices checked November 3, 2026.

Claude Opus 5: price, context window and benchmarks

Claude Opus 5 is Anthropic's previous Opus model, released in July 2026 at $5 per million input tokens and $25 per million output tokens. Anthropic now lists it as legacy: Claude Opus 5.5 costs less and scores higher on every benchmark both models have a score for, so new projects should start there.

Best for

Key facts

FactClaude Opus 5Source
ProviderAnthropicClaude docs: Pricing
API model nameclaude-opus-5Claude docs: Claude Opus 5
ReleasedJuly 24, 2026Anthropic: Introducing Claude Opus 5
Input price$5 per million tokensClaude docs: Pricing
Cached input price$0.50 per million tokensClaude docs: Pricing
Output price$25 per million tokensClaude docs: Pricing
Context window1,000,000 tokensClaude docs: Claude Opus 5
Maximum output128,000 tokensClaude docs: Claude Opus 5
Knowledge cutoffMay 2026Claude docs: Claude Opus 5
Input and outputText and images in; text outClaude docs: Claude Opus 5
ReasoningAdaptive thinking on by default; it can be switched off at high effort or below. Effort levels low to max; high is the API defaultClaude docs: What's new in Claude Opus 5.5 (Opus 5 thinking rules)
API and cloud platformsClaude API, Amazon Bedrock, Google Cloud, Microsoft Foundry and Claude Platform on AWS; legacy model, retirement no sooner than July 24, 2027Claude docs: Claude Opus 5

Cache writes cost $6.25 per million tokens (5-minute cache) or $10 (1-hour cache). The Batch API costs half. There is no long-context surcharge. Fast mode costs $10 in and $50 out.

What Claude Opus 5 costs a month

Three typical workloads, priced at the standard API rates above. We calculated these; the provider does not publish them.

WorkloadTokens a monthCost a month
Support chatbot
100,000 replies, each 3,000 input and 400 output tokens, no caching
300M in, 40M out $2,500
Document Q&A
5,000 questions over the same 60,000-token document set (90% cache hits), 800 output tokens each
300M in, 4M out $385
Coding agent
200 tasks, each 1.5 million input tokens (90% cache hits) and 50,000 output tokens
300M in, 10M out $535

Reasoning models bill their thinking as output tokens, so heavy reasoning can raise the output count several times. Cache-write charges are not included.

Benchmarks

BenchmarkScoreSource
Terminal-Bench 4.0
max effort
49.0% GPT-6 Astra vs Claude Opus 5
Terminal-Bench-Science 0.1
max effort
29.0% GPT-6 Astra vs Claude Opus 5

Speed: about 49.5 output tokens a second (high effort, Artificial Analysis; Claude Opus 5 (high)).

Benchmark and speed data marked as such come from Artificial Analysis.

What changed

Opus 5 kept Opus 4.8's price of $5 and $25 per million tokens. Anthropic said at launch that it came close to Claude Fable 5 at half the price. Claude Opus 5.5 replaced it as the recommended Opus on September 22, 2026. (Anthropic: Introducing Claude Opus 5)

Claude Opus 5 compared

Other models

All prices side by side: AI model API price table. How we research: our method.

Sources