o3 (2025-04-16)
Published by OpenAI. This profile currently contains 3 independently sourced benchmark results.
View model profile →Evidence-based comparison
o3 (2025-04-16) vs o4-mini (2025-04-16)
Compare the configurations using only independent benchmarks where both have published results under the same version and methodology.
Like-for-like results
Shared benchmark scores
No cross-benchmark averaging. A higher score is better for every comparison shown below.
Benchmarko3 (2025-04-16)o4-mini (2025-04-16)
HELM Capabilities0.8110.812
HELM Safety0.9820.973
Arena-Hard v2.0 Preview85.9%74.6%
o4-mini (2025-04-16)
Published by OpenAI. This profile currently contains 4 independently sourced benchmark results.
View model profile →The comparison that affects your business
Which model recommends you?
Measure your brand across the AI assistants customers use to discover and compare businesses.
Run my free visibility check →