AI Benchmark Model

GPT-5.3 Codex (xhigh)

Benchmark scores, pricing, speed, and model comparisons for GPT-5.3 Codex (xhigh).

Key scores

Intelligence53.6Index score
Coding53.1Index score
Speed71.1Tokens/sec
Blended Cost$4.81Per 1M tokens

The GPT-5.3 Codex (xhigh) model, developed by OpenAI and released on February 5, 2026, is a specialized artificial intelligence system designed for high-performance coding and reasoning tasks. The model holds an intelligence index of 53.6 and a coding index of 53.1. Performance metrics indicate an output speed of 92.337 tokens per second with a time to first token of 54.938 seconds. Pricing for the model is set at $1.75 per million input tokens and $14.00 per million output tokens, resulting in a blended rate of $4.81 per million tokens. Benchmark results include a GPQA score of 0.915, HLE of 0.399, SciCode of 0.532, IFBench of 0.754, LCR of 0.74, TerminalBench Hard of 0.530, and TAU2 of 0.860.

Benchmark results

7 tracked benchmark rows for GPT-5.3 Codex (xhigh).

GPQA Reasoning
91.5%
TAU2 Other
86.0%
IFBench Other
75.4%
LCR Other
74.0%
SciCode Other
53.2%
TerminalBench Hard Other
53.0%
HLE Other
39.9%

Comments (0)

No comments yet

Be the first to share your thoughts!