Context
n/a
Max output
n/a
Serving providers
1
Cheapest input
$3.00 /1M
On-demand listed rates, OpenRouter routed quotes and Vercel reseller quotes. Enterprise commitments (committed-use discounts, provisioned throughput, negotiated rates) can be materially lower.
Benchmarks
independent evaluations · source: Artificial AnalysisSpeed & latency
Output speed
42 t/s
tokens / second
Time to first token
1.01s
latency
Headline indices
Individual benchmarks · click a row for effort variants
| Benchmark | Best score | Effort | Cost / task | |||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Automation Bench | 0.1 | reasoning: true | — | |||||||||||||||||
| CritPt | 1% | reasoning: true | — | |||||||||||||||||
| ||||||||||||||||||||
| GDPval v2Frozen Aug 18, 2026 | 1055.5 | reasoning: true | — | |||||||||||||||||
| ||||||||||||||||||||
| GPQA Diamond | 83% | reasoning: true | — | |||||||||||||||||
| ||||||||||||||||||||
| Humanity's Last Exam | 18% | reasoning: true | — | |||||||||||||||||
| ||||||||||||||||||||
| IFBench | 57% | reasoning: true | — | |||||||||||||||||
| ||||||||||||||||||||
| LiveCodeBench | 71% | reasoning: true | — | |||||||||||||||||
| ||||||||||||||||||||
| Long-Context Reasoning | 72% | reasoning: true | — | |||||||||||||||||
| ||||||||||||||||||||
| MMLU-ProFrozen Aug 18, 2026 | 87% | reasoning: true | — | |||||||||||||||||
| ||||||||||||||||||||
| MMMU-Pro | 69% | reasoning: true | — | |||||||||||||||||
| ||||||||||||||||||||
| Omniscience | -0.1 | reasoning: true | — | |||||||||||||||||
| ||||||||||||||||||||
| Omniscience Accuracy | 0.3 | reasoning: true | — | |||||||||||||||||
| ||||||||||||||||||||
| Omniscience Non Hallucination | 0.5 | reasoning: true | — | |||||||||||||||||
| ||||||||||||||||||||
| SciCode | 46% | reasoning: true | — | |||||||||||||||||
| ||||||||||||||||||||
| Terminal Bench40 | 0 | reasoning: true | — | |||||||||||||||||
| ||||||||||||||||||||
| TerminalBench Hard | 36% | reasoning: true | — | |||||||||||||||||
| ||||||||||||||||||||
| TerminalBench v2.1 | 56% | reasoning: true | — | |||||||||||||||||
| ||||||||||||||||||||
| τ-bench Banking | 25% | reasoning: true | — | |||||||||||||||||
| τ²-bench | 78% | reasoning: true | — | |||||||||||||||||
| ||||||||||||||||||||
Artificial Analysis last published scores for this benchmark on Sep 9, 2026. Scores here are frozen at that date, and models released since are not on it.
Artificial Analysis last published scores for this benchmark on Aug 18, 2026. Scores here are frozen at that date, and models released since are not on it.
Pricing by serving provider
standard tier · per 1M tokens · click a provider to expand| Serving provider | Input /1M | Output /1M | Endpoints | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ReplicateCheapest | $3.00 | $15.00 | 1 | ||||||||||
Standardreplicate/anthropic/claude-4.5-sonnet
| |||||||||||||
Price history
listed seller prices, input and output per 1M tokens Input Output
Cost calculator
estimate your monthly spend on this model$1,800
estimated / month
Model IDs
copy the exact identifier for your platformreplicate/anthropic/claude-4.5-sonnet
Frequently asked questions
Claude 4.5 Sonnet pricing, context and availabilityHow much does Claude 4.5 Sonnet cost?
Claude 4.5 Sonnet costs $3.00 per 1M input tokens from Replicate, and $15.00 per 1M output tokens from Replicate.
What can Claude 4.5 Sonnet do?
Claude 4.5 Sonnet supports image input (vision), tool use, prompt caching and structured output.
Other Anthropic models
compare pricing across the Anthropic lineupClaude Haiku 4.5$1.00/1M · 11 providersClaude Sonnet 5$2.00/1M · 9 providersClaude Sonnet 4.6$3.00/1M · 9 providersClaude Opus 5$5.00/1M · 9 providersClaude Fable 5$10.00/1M · 9 providersClaude Sonnet 4.5$3.00/1M · 8 providersClaude Opus 4.7$5.00/1M · 8 providersClaude Opus 4.8$5.00/1M · 8 providers
Listed prices from LiteLLM, OpenRouter and Vercel · last checked Oct 14, 2026
Every weekday
AI moves fast. Here's your debrief.
News, analysis, tools, and more.
For people who build with AI