Analysis of Performance Gap Between Open Source and Closed Source LLMs

Jun 26, 2026

AI Summary

A recent analysis examines the performance gap between open source and closed source large language models (LLMs) using various benchmarks. While the gap appears to be closing, particularly in coding capabilities, the overall average suggests that open source models remain about five months behind their closed source counterparts.

The analysis measures the performance gap between open weights LLMs and closed source LLMs using the Artificial Analysis Intelligence Index benchmark.
The gap has been shrinking since summer 2024 and is projected to reach zero by December 3, 2026, based on a line of best fit.
However, this analysis is based on a single benchmark, and further examination across 18 different benchmarks shows that the average gap remains just under five months.
Significant improvements have been noted in the coding benchmark, where the gap has reduced from 15 months to about one or two months.
The findings highlight the complexity of measuring LLM quality, with results varying depending on the benchmark used.

open weightsclosed sourcellmsai researchmachine learning

Analysis of Performance Gap Between Open Source and Closed Source LLMs

Related Stories

OpenAI says GPT 5.6 is the ‘preferred model’ for Microsoft Copilot amid breakup chatter

OpenAI introduces GPT-5.6 models with enhanced efficiency and cybersecurity features

OpenAI Releases GPT-5.6 Model with Enhanced Capabilities

OpenAI launches GPT-5.6 Sol, improving token efficiency for coding tasks