Claude Sonnet 4.5 vs GPT-4o

Benchmarks, API pricing and specs, head to head. Data updated 2026-06-10.

Claude Sonnet 4.5

Anthropic · Sep 2025

65.5

The model that made 30-hour autonomous coding sessions real — 77.2% SWE-bench Verified and a 1M-token context beta at $3/$15.

GPT-4o

OpenAI · May 2024

The 2024 omni-modal classic with native audio in and out; superseded at the frontier but still everywhere in production.

The verdict

Claude Sonnet 4.5 wins 3 of the 3 benchmarks these models share, against 0 for GPT-4o. GPT-4o is about 1.4x cheaper per blended million tokens (3:1 input:output mix). Claude Sonnet 4.5 also takes 200K of context versus 128K for GPT-4o.

Benchmark head-to-head 30

SWE-bench Verified
77.2%33.2%
Claude Sonnet 4.5GPT-4o
GPQA Diamond
83.4%53.6%
Claude Sonnet 4.5GPT-4o
MMMU
~77.8%69.1%
Claude Sonnet 4.5GPT-4o

Specs & pricing

Claude Sonnet 4.5GPT-4o
modhub Index65.5
Input price / 1M$3$2.5
Output price / 1M$15$10
Context window200K128K
Max output64K16K
Open weightsnono
Reasoning modelyesno
Multimodal inputtext, imagetext, image, audio
Knowledge cutoffJan 2025Oct 2023
ReleasedSep 2025May 2024
Example monthly cost*$52.50$40.00

* 10M input + 1.5M output tokens per month at list prices, no caching. Green = better value on that row.

Frequently asked questions

Which is better, Claude Sonnet 4.5 or GPT-4o?
Claude Sonnet 4.5 wins 3 of the 3 benchmarks these models share, against 0 for GPT-4o. GPT-4o is about 1.4x cheaper per blended million tokens (3:1 input:output mix). Claude Sonnet 4.5 also takes 200K of context versus 128K for GPT-4o.
Which is cheaper, Claude Sonnet 4.5 or GPT-4o?
Claude Sonnet 4.5 costs $3/$15 per million input/output tokens, while GPT-4o costs $2.5/$10. For a typical workload of 10M input and 1.5M output tokens per month, that's $52.50 versus $40.00.
Which model is better for coding, Claude Sonnet 4.5 or GPT-4o?
On SWE-bench Verified — the standard agentic-coding benchmark — Claude Sonnet 4.5 scores 77.2% versus 33.2% for GPT-4o, making Claude Sonnet 4.5 the stronger pick for coding agents.

More comparisons