▸ compare
OpenAI: GPT-5.4 vs Anthropic: Claude Sonnet 4.5
Side-by-side benchmark comparison based on independent public data. Scores, pricing, context, and task breakdown.
▸ verdict
OpenAI: GPT-5.4
90.2
higher score
vs
Anthropic: Claude Sonnet 4.5
85.7
OpenAI: GPT-5.4 leads with a higher composite benchmark score. It's the stronger choice for general-purpose tasks based on public benchmark data.
▸ score breakdown
| Task | OpenAI: GPT-5.4 | Anthropic: Claude Sonnet 4.5 | Δ |
|---|---|---|---|
| Overall | 90.2 | 85.7 | +4.5 |
| Coding | — | 59.8 | — |
| Reasoning | 74.4 | 45.3 | +29.1 |
| Math | — | — | — |
| Writing | 88.9 | 84.3 | +4.6 |
| JSON | — | 28.6 | — |
▸ specs & pricing
| Attribute | OpenAI: GPT-5.4 | Anthropic: Claude Sonnet 4.5 |
|---|---|---|
| Provider | Openai | Anthropic |
| Context window | 1M | 1M |
| Input $/M tokens | $2.50/M | $3.00/M |
| Output $/M tokens | $15.00/M | $15.00/M |
| Weights | proprietary | proprietary |
▸ frequently asked
Is OpenAI: GPT-5.4 better than Anthropic: Claude Sonnet 4.5?
OpenAI: GPT-5.4 scores higher overall (90.2 vs 85.7) in the benchmark composite. The best choice depends on the specific use case and budget.
Which is cheaper: OpenAI: GPT-5.4 or Anthropic: Claude Sonnet 4.5?
OpenAI: GPT-5.4 is cheaper at $2.5/M input tokens vs $3/M for Anthropic: Claude Sonnet 4.5.
Which model is better for coding?
Both models perform similarly on coding benchmarks (SWE-Bench, Aider Polyglot, LiveCodeBench).