GPT-5
ActiveOpenAI's frontier flagship model with long context and advanced reasoning.
Overview
GPT-5 is OpenAI's top-tier general-purpose model with strong reasoning, coding, and multimodal capabilities. It supports tool calling, structured outputs, and long context.
Benchmarks
| Benchmark | Score | Source |
|---|---|---|
| AIME 2024Math | 60% accuracy | Self-reported OpenAI system card |
| GPQA DiamondReasoning | 72% accuracy | Third-party Artificial Analysis |
| GSM8KMath | 96.1% accuracy | Self-reported OpenAI system card |
| HumanEvalCoding | 94pass@1 % | Self-reported OpenAI system card |
| MATHMath | 87.1% accuracy | Self-reported OpenAI system card |
| MMLUGeneral knowledge | 88% accuracy | Self-reported OpenAI system card |
| MMLU-ProGeneral knowledge | 76% accuracy | Third-party Artificial Analysis |
| SWE-bench VerifiedCoding | 48% resolved | Third-party Third-party harness run |
Integrations & tooling support
- Tool calling
- Supported
- Structured outputs
- Supported
Price vs quality
Fair value
Mid-tier performance and pricing — standard choice.
- Quality percentile
- 50.8%
- Effective price
- $7.8125/1M
- Pricing breakdown
- $1.25/1M in
$10/1M out
vs 8 benchmarks
/ 1M tokens (input + 3× output)
Community ratings
No ratings yet. Be the first to rate GPT-5.
Rate GPT-5
Sign in to rate and review.
Comments
Sign in to leave a comment.