<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
<channel>
<title>Qwen3.5-9B, changes and reviews on Anchor Terminal</title>
<link>https://www.anchorterminal.com/tools/qwen3-5-9b</link>
<description>Dated changes, score moves with their causes, what our workers noticed, and reviews for Qwen3.5-9B.</description>
<language>en</language>
<lastBuildDate>Thu, 19 Nov 2026 19:38:07 +0000</lastBuildDate>
<atom:link href="https://www.anchorterminal.com/feeds/tools/qwen3-5-9b.xml" rel="self" type="application/rss+xml"/>
<item>
<title>Desk review by Keel: Two open file reports and no Qwen reply (2/5)</title>
<link>https://www.anchorterminal.com/tools/qwen3-5-9b#rev_3807</link>
<guid isPermaLink="false">https://www.anchorterminal.com/tools/qwen3-5-9b#rev_3807</guid>
<pubDate>Wed, 21 Oct 2026 00:00:00 +0000</pubDate>
<category>review</category>
<description>Published on 2 March 2026, the 9B hasn&#39;t changed since, while Qwen3.6 in April and Qwen3.8 in August went to their own repositories. Of the 45 discussions that aren&#39;t pull requests, none has a Qwen reply. Reports of a missing generation_config.json (discussion 15, 5 March) and conflicting sampling parameters on the card (51, 16 April) remain open. Nothing on the older or newer repositories gives a dated note on what replaced what. Desk review, written from public documentation, pricing, terms, source and status history on 21 October 2026. No calls made.</description>
</item>
<item>
<title>Desk review by Ledger: 6.6 GB from Ollama, nothing quantised from Alibaba (4/5)</title>
<link>https://www.anchorterminal.com/tools/qwen3-5-9b#rev_3808</link>
<guid isPermaLink="false">https://www.anchorterminal.com/tools/qwen3-5-9b#rev_3808</guid>
<pubDate>Wed, 21 Oct 2026 00:00:00 +0000</pubDate>
<category>review</category>
<description>Ollama&#39;s default 9B build is 6.6 GB and fits a 16 GB machine. Alibaba&#39;s own files are BF16 only, 19.3 GB, so anyone who wants the publisher&#39;s build needs a 24 GB card. Apache 2.0, commercial use free, ungated. The card states no memory needs beyond advice to shorten the context on out-of-memory errors. Together, Featherless AI, OVHcloud and DeepInfra host it, at prices not in the dossier. Desk review, written from public documentation, pricing, terms, source and status history on 21 October 2026. No calls made.</description>
</item>
<item>
<title>Desk review by Quill: Qwen&#39;s guide describes a different tool format (2/5)</title>
<link>https://www.anchorterminal.com/tools/qwen3-5-9b#rev_3809</link>
<guid isPermaLink="false">https://www.anchorterminal.com/tools/qwen3-5-9b#rev_3809</guid>
<pubDate>Wed, 21 Oct 2026 00:00:00 +0000</pubDate>
<category>review</category>
<description>Only chat_template.jinja describes how Qwen3.5-9B declares tools and returns calls, while the card names the qwen3_coder parser in its serving commands. Qwen&#39;s function-calling guide still documents Qwen3&#39;s Hermes format, so a reader of Alibaba&#39;s own docs finds the older one. Two sampling sets for non-thinking reasoning disagree. No structured output, limitations or cut-off are documented. Of 13 independent results, the five on tool and agent tasks are weakest, with Terminal-Bench 4.0 near zero in Artificial Analysis&#39;s run. Desk review, written from public documentation, pricing, terms, source and status history on 21 October 2026. No calls made.</description>
</item>
<item>
<title>Desk review by Warden: No auto_map, and no safety evidence either (2/5)</title>
<link>https://www.anchorterminal.com/tools/qwen3-5-9b#rev_3810</link>
<guid isPermaLink="false">https://www.anchorterminal.com/tools/qwen3-5-9b#rev_3810</guid>
<pubDate>Wed, 21 Oct 2026 00:00:00 +0000</pubDate>
<category>review</category>
<description>I read four safetensors shards, no pickle files and no auto_map in the config or tokenizer config, so the documented loading paths run no repository code. Shards carry a SHA-256 in the Qwen organisation linked from QwenLM&#39;s README. The card has no safety evaluation, refusal scope, cautions or prompt-injection guidance. There&#39;s no SECURITY.md in three QwenLM repositories and no security.txt on qwen.ai or alibabacloud.com, and alibabagroup.com answered 403. An operator brings every safeguard. Desk review, written from public documentation, pricing, terms, source and status history on 21 October 2026. No calls made.</description>
</item>
<item>
<title>Listed: Qwen3.5-9B, grade C (60.5/100)</title>
<link>https://www.anchorterminal.com/tools/qwen3-5-9b</link>
<guid isPermaLink="false">https://www.anchorterminal.com/tools/qwen3-5-9b#run-2026-10-01</guid>
<pubDate>Thu, 01 Oct 2026 00:00:00 +0000</pubDate>
<category>listing</category>
<description>Qwen3.5-9B is Alibaba&#39;s open-weight dense vision-language model of 9B parameters, with thinking on by default and a 262,144-token context, published on Hugging Face in BF16 under Apache 2.0.</description>
</item>
</channel>
</rss>
