Compare AI models

Specifications
Parameters
—4B/30B/235B
Context window
128k256k
These models have no shared benchmark scores.
Overview
CompanyOpenAIQwen
Release dateFeb 12 2026Jul 21 2025
AccessClosedOpen Weight
Model detailsView modelView model

Frequently asked questions

GPT-5.3-Codex-Spark and Qwen3-2507 don't publish scores on any of the same benchmarks, so there's no direct head-to-head comparison. Qwen3-2507 shipped 206 days before GPT-5.3-Codex-Spark, so benchmark comparisons should account for the intervening progress.

Context windows are 128k (GPT-5.3-Codex-Spark) vs 256k (Qwen3-2507). GPT-5.3-Codex-Spark is closed, while Qwen3-2507 is open weight.

Direct benchmark comparisons are unavailable — GPT-5.3-Codex-Spark and Qwen3-2507 don't publish scores on any of the same benchmarks.