GPT-5.4 mini
Current
- Context window
- 400.000 tokens
- Max output
- 16.384 tokens
- Price input / 1M tokens
- $0.75
- Price output / 1M tokens
- $4.50
- Price cache-read / 1M
- $0.08
Capabilities
- Vision
- Thinking
- Tool use
- Cache
Benchmark scores
- GPQA Diamond
- 87.5
- SWE-bench Pro
- 54.4
- OSWorld-Verified
- 72.1
- Terminal-Bench 2.0
- 60.0
Notes
Compact GPT-5.4 variant, competitively priced. Source: openai.com/api/pricing.