OpenAI

o3

Multimodal
Zero-eval
#1ARC-AGI
#1MathVista
#1Tau2 Airline
+13 more

by OpenAI

+
+
+
+
About

o3 represents the next generation in OpenAI's reasoning model series, developed to advance the capabilities of deliberate, step-by-step problem solving. Built to handle increasingly complex challenges across mathematics, science, and coding, it continues the evolution of reasoning-focused AI with improved analytical depth and accuracy.

+
+
+
+
Pricing Range
Input (per 1M)$2.00 -$2.00
Output (per 1M)$8.00 -$8.00
Providers1
+
+
+
+
Timeline
AnnouncedApr 16, 2025
ReleasedApr 16, 2025
Knowledge CutoffMay 31, 2024
+
+
+
+
Specifications
Capabilities
Multimodal
+
+
+
+
License & Family
License
Proprietary
Performance Overview
Performance metrics and category breakdown

Overall Performance

22 benchmarks
Average Score
67.5%
Best Score
98.4%
High Performers (80%+)
10

Performance Metrics

Max Context Window
300.0K
Avg Throughput
50.0 tok/s
Avg Latency
20ms
+
+
+
+
All Benchmark Results for o3
Complete list of benchmark scores with detailed information
COLLIE
text
0.98
98.4%
Self-reported
AIME 2024
text
0.92
91.6%
Self-reported
ARC-AGI
image
0.88
88.0%
Self-reported
MathVista
multimodal
0.87
86.8%
Self-reported
AIME 2025
text
0.86
86.4%
Self-reported
VideoMMMU
multimodal
0.83
83.3%
Self-reported
GPQA
text
0.83
83.3%
Self-reported
MMMU
multimodal
0.83
82.9%
Self-reported
Aider-Polyglot
text
0.81
81.3%
Self-reported
Tau2 Retail
text
0.80
80.2%
Self-reported
Showing 1 to 10 of 22 benchmarks
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+