---
title: "Qwen 3.8 27B Scores 52, Matching GPT-5.6 Luna on AI Index"
source: "Simon Willison 博客（Atom）"
source_tier: "t2"
url: "https://simonwillison.net/2026/Aug/17/qwen-38-27b-scores-52/"
permalink: "https://feed.astralor.com/items/itm_6aa6db5f712edc5a53ae85a7"
published: "2026-08-17T23:58:14.000Z"
score: 69
channel: "ai"
tags: ["Artificial Analysis Intelligence Index","DeepSeek-V4-Pro","GLM-5.2","GPT-5.6 Luna","Qwen 3.8","Open-source Model"]
---

# Qwen 3.8 27B Scores 52, Matching GPT-5.6 Luna on AI Index

## AI commentary

> The most interesting part is not the score itself, but that a 27B model lands within one point of a 753B model. For teams building agents or doing private deployments, this means top-tier capability can fit in much smaller memory budgets, dramatically improving cost-effectiveness. As always, single-index scores should be validated on real tasks.

## At a glance

Open-source model Qwen 3.8 27B scored 52 on the Artificial Analysis Intelligence Index, the same as GPT-5.6 Luna (max) and just one point behind GLM-5.2 (max) and DeepSeek V4 Pro 0813 (max). Given that GLM-5.2 has 753B parameters and DeepSeek V4 Pro is also a large-scale model, Qwen 3.8 27B achieves near-top performance with only 27B parameters, highlighting its efficiency and cost advantages for local deployment and agent workloads.

## Key points

- Qwen 3.8 27B scored 52 on the Artificial Analysis Intelligence Index.
- It matches GPT-5.6 Luna (max) and trails GLM-5.2 (max) and DeepSeek V4 Pro 0813 (max) by just one point.
- With only 27B parameters, it is far smaller than GLM-5.2's 753B, showing a clear efficiency advantage.
- This demonstrates that small open-source models can approach the capability level of top-tier large models.

## Full text

That's the same score as GPT-5.6 Luna (max), and just one point behind GLM-5.2 (max) and DeepSeek V4 Pro 0813 (max) - that GLM is 753B and that DeepSeek is …

---

Verify against the original before citing; curated by Astral Feed: https://feed.astralor.com/items/itm_6aa6db5f712edc5a53ae85a7