GPT-4o vs Claude Sonnet 4.6 API Price Comparison
Prices updated: ·137 quotes
As of : At official rates, GPT-4o is $2.5000/1M input tokens and Claude Sonnet 4.6 is $3.0000 — GPT-4o is ~17% cheaper than Claude Sonnet 4.6. Via relay providers, the lowest GPT-4o rate is $0.0828 (PoloAPI) and Claude Sonnet 4.6 is $0.0190 (OmniTunnel). Context windows: GPT-4o 131K, Claude Sonnet 4.6 1024K. GPT-4o and Claude Sonnet 4.6 support image input. 137 provider quotes tracked (GPT-4o: 100, Claude Sonnet 4.6: 37), updated daily.
GPT-4oMid-tier
Official
$2.5000
/ 1M tokens
Best Price
$0.0828
/ 1M tokens
100 providers · 131K Context · 👁 Vision
Claude Sonnet 4.6Mid-tier
Official
$3.0000
/ 1M tokens
Best Price
$0.0190
/ 1M tokens
37 providers · 1024K Context · 👁 Vision
At official prices, GPT-4o is ~17% cheaper. At the best relay price, Claude Sonnet 4.6 is ~77% cheaper.
Price Comparison
| GPT-4o | Claude Sonnet 4.6 | |
|---|---|---|
| Official input / 1M tokens | $2.5000 ✓ | $3.0000 |
| Official output / 1M tokens | $10.0000 ✓ | $15.0000 |
| Cheapest input / 1M tokens | $0.0828 | $0.0190 ✓ |
| Cheapest output / 1M tokens | $0.0517 ✓ | $0.0970 |
| providers | 100 | 37 |
| Context | 131K | 1024K |
| Vision | ✓ | ✓ |
GPT-4o — All Providers
Detail page →Claude Sonnet 4.6 — All Providers
Detail page →Frequently Asked Questions
Which is cheaper, GPT-4o or Claude Sonnet 4.6?
At official price, GPT-4o is cheaper by ~17%. At the best relay price, Claude Sonnet 4.6 with a ~77% difference.
What context window do GPT-4o and Claude Sonnet 4.6 support?
GPT-4o has a 131K token context window; Claude Sonnet 4.6 has 1024K tokens.
Do GPT-4o and Claude Sonnet 4.6 support image input?
GPT-4o supports image input; Claude Sonnet 4.6 supports image input.
What is the cheapest way to access GPT-4o or Claude Sonnet 4.6?
GPT-4o's lowest input price is $0.0828/1M tokens (PoloAPI). Claude Sonnet 4.6's is $0.0190/1M tokens (OmniTunnel). Relay providers typically offer lower prices with different SLA terms.
Should developers choose GPT-4o or Claude Sonnet 4.6?
Test both with representative prompts and output lengths, then compare quality, total official or relay cost, context, latency, and SLA. The lowest token rate may not produce the lowest project cost.
Related Comparisons
Related