Requesty
Anthropic PBC logo
Anthropic PBC

claude-sonnet-4-5

Claude Sonnet 4.5 is Anthropic's best coding model in the world, leading on SWE-bench Verified (77.2%) and OSWorld (61.4%). It delivers sustained autonomous performance on complex tasks for over 30 hours—up from seven hours for Opus 4—maintaining focus and reliability throughout the entire software development lifecycle, with enhanced capabilities in tool handling, memory management, and context processing that make it the strongest model for building complex agents.

👁Vision🧠Reasoning🔧Tool calling⚡Caching🖥Computer use
Try claude-sonnet-4-5CompareDocs

Specifications

Context window1M tokens
Max output64K tokens
API typechat
AddedSep 29, 2025
Model IDanthropic/claude-sonnet-4-5
Data retentionYes (30 days)
Used for trainingNo
Provider location🇺🇸 US
Privacy policyAnthropic Privacy Policy →

Benchmarks

Released 2025-09
SWE-Bench Verifiedcoding
70.8%

Resolving real GitHub issues from 12 popular Python repositories.

GPQA Diamondreasoning
74.2%

Graduate-level physics, chemistry & biology questions designed to resist Googling.

MMLU Proknowledge
83.8%

Massive Multitask Language Understanding across 57 academic subjects.

Scores are sourced from official model cards, Artificial Analysis, and public leaderboards. Benchmarks measure specific skills and do not capture every aspect of model quality — always test on your own workload.

Pricing

Input / 1M
$3.00
Output / 1M
$15.00
Cache write / 1M
$3.75
Cache read / 1M
$0.30
Estimated cost
100K input + 10K output$0.45
1M input + 100K output$4.50
10M input + 1M output$45.00

Requesty charges exactly what the upstream provider charges — no markup, no per-request fees. Prompt caching and smart routing can reduce effective cost by 30-80%.

Quickstart

Drop-in compatible with the OpenAI SDK. Change the base URL, swap in your Requesty API key, and set the model to anthropic/claude-sonnet-4-5.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
from openai import OpenAI

client = OpenAI(
    api_key="YOUR_REQUESTY_API_KEY",
    base_url="https://router.requesty.ai/v1",
)

response = client.chat.completions.create(
    model="anthropic/claude-sonnet-4-5",
    messages=[
        {"role": "user", "content": "Explain quantum computing in one paragraph."},
    ],
)

print(response.choices[0].message.content)

Other Anthropic PBC models

claude-opus-4-7
1M ctx$5.00/1M
SWE-Bench 79%
claude-sonnet-4-6
1M ctx$3.00/1M
SWE-Bench 77%
claude-opus-4-6
1M ctx$5.00/1M
SWE-Bench 75%
claude-haiku-4-5
200K ctx$1.00/1M
SWE-Bench 54%
claude-opus-4-5
200K ctx$5.00/1M
SWE-Bench 71%
claude-opus-4-1
200K ctx$15.00/1M

Frequently asked questions

How much does claude-sonnet-4-5 cost?
claude-sonnet-4-5 is priced at $3.00 per million input tokens and $15.00 per million output tokens when accessed via Requesty. Prompt caching is supported, which can cut effective input cost by up to 90% on repeated context. Requesty charges exactly what the upstream provider charges — we don't add markup.
What is the context window of claude-sonnet-4-5?
claude-sonnet-4-5 has a context window of 1M tokens, with a maximum output of 64K tokens per response. That's roughly 1,333 words of input you can fit in a single prompt.
How does claude-sonnet-4-5 perform on benchmarks?
claude-sonnet-4-5 scores 92.0% on HumanEval, 87.5% on MATH, 83.8% on MMLU Pro. See the full benchmark chart above for results across MMLU Pro, GPQA Diamond, SWE-Bench Verified, HumanEval, MATH, AIME, MMMU, and LiveBench.
What can claude-sonnet-4-5 do?
claude-sonnet-4-5 supports vision input, tool calling, extended reasoning, prompt caching, computer use. You can call it through any OpenAI-compatible client by pointing base_url to Requesty.
How do I use claude-sonnet-4-5 with the OpenAI SDK?
Install the OpenAI SDK, set base_url to "https://router.requesty.ai/v1", set your API key to your Requesty key, and set the model to "anthropic/claude-sonnet-4-5". The Quickstart above shows Python, JavaScript and cURL snippets.

Access claude-sonnet-4-5 through Requesty

One API key, 400+ models, OpenAI-compatible. No markup on provider prices, automatic failover, and smart caching built-in.

Get started freeAll Anthropic PBC models