{"id":1783,"date":"2026-04-05T08:30:09","date_gmt":"2026-04-05T08:30:09","guid":{"rendered":"https:\/\/wiro.ai\/blog\/?p=1783"},"modified":"2026-03-19T10:24:52","modified_gmt":"2026-03-19T10:24:52","slug":"seed-v2-mini-vs-qwen3-5-27b-5-small-tests","status":"publish","type":"post","link":"https:\/\/wiro.ai\/blog\/seed-v2-mini-vs-qwen3-5-27b-5-small-tests\/","title":{"rendered":"Seed-V2 Mini vs Qwen3.5-27B: 5 Small Tests"},"content":{"rendered":"<p>Seed-V2 Mini vs Qwen3.5-27B sounds like a simple comparison. The outputs can look very different in practice. This post runs five small tests with the same prompts and shows the raw results.<\/p>\n<h2>Models<\/h2>\n<ul>\n<li><a href=\"https:\/\/wiro.ai\/models\/bytedance\/seed-v2-mini\">Seed-V2 Mini on Wiro<\/a><\/li>\n<li><a href=\"https:\/\/wiro.ai\/models\/qwen\/qwen3-5-27b\">Qwen3.5-27B on Wiro<\/a><\/li>\n<\/ul>\n<h2>Quick take<\/h2>\n<ul>\n<li>Seed-V2 Mini returned cleaner, direct outputs on these prompts.<\/li>\n<li>Qwen3.5-27B often wrapped answers with a thinking-style format unless the client disables it.<\/li>\n<\/ul>\n<h2>Results (same prompt, two outputs)<\/h2>\n<h3>Test 1: JSON extraction<\/h3>\n<p><strong>Prompt<\/strong><\/p>\n<pre><code>Extract JSON with keys name, email, plan, seats. Return JSON only.\n\nInput: Name: Jamie Lee; Email: jamie@example.com; Plan: Pro Annual; Seats: 12<\/code><\/pre>\n<p><strong>Seed-V2 Mini<\/strong><\/p>\n<pre><code>{\n  \"name\": \"Jamie Lee\",\n  \"email\": \"jamie@example.com\",\n  \"plan\": \"Pro Annual\",\n  \"seats\": 12\n}<\/code><\/pre>\n<p><strong>Qwen3.5-27B<\/strong><\/p>\n<pre><code>{\n \"name\": \"Jamie Lee\",\n \"email\": \"jamie@example.com\",\n \"plan\": \"Pro Annual\",\n \"seats\": 12\n }<\/code><\/pre>\n<h3>Test 2: JavaScript utility function<\/h3>\n<p><strong>Prompt<\/strong><\/p>\n<pre><code>Write a JavaScript function slugify(str) that lowercases, replaces spaces with hyphens, removes non alphanumeric characters (keep hyphens), collapses multiple hyphens, and trims leading and trailing hyphens. Return code only.<\/code><\/pre>\n<p><strong>Seed-V2 Mini<\/strong><\/p>\n<pre><code>function slugify(str) {   return str     .toLowerCase()     .replace(\/[^a-z0-9-]\/g, '')     .replace(\/\\s+\/g, '-')     .replace(\/-+\/g, '-')     .replace(\/^-|-$\/g, ''); }<\/code><\/pre>\n<p><strong>Qwen3.5-27B<\/strong><\/p>\n<pre><code>function slugify(str) {\n return str\n .toLowerCase()\n .replace(\/\\s+\/g, '-')\n .replace(\/[^a-z0-9-]\/g, '')\n .replace(\/-+\/g, '-')\n .replace(\/^-+|-+$\/g, '');\n }<\/code><\/pre>\n<h3>Test 3: English to Turkish<\/h3>\n<p><strong>Prompt<\/strong><\/p>\n<pre><code>Translate to Turkish. Return Turkish only.\n\nText: This model returns thinking text by default. Disable it for clean JSON outputs.<\/code><\/pre>\n<p><strong>Seed-V2 Mini<\/strong><\/p>\n<pre><code>Bu model varsay\u0131lan olarak d\u00fc\u015f\u00fcnme metni d\u00f6nd\u00fcr\u00fcr. Temiz JSON \u00e7\u0131kt\u0131lar\u0131 i\u00e7in bunu devre d\u0131\u015f\u0131 b\u0131rak\u0131n.<\/code><\/pre>\n<p><strong>Qwen3.5-27B<\/strong><\/p>\n<pre><code>Bu model varsay\u0131lan olarak d\u00fc\u015f\u00fcnme metni d\u00f6nd\u00fcr\u00fcr. Temiz JSON \u00e7\u0131kt\u0131lar\u0131 i\u00e7in devre d\u0131\u015f\u0131 b\u0131rak\u0131n.<\/code><\/pre>\n<h3>Test 4: 5-bullet summary<\/h3>\n<p><strong>Prompt<\/strong><\/p>\n<pre><code>Summarize in 5 bullet points. Return bullets only.\n\nText: Wiro runs models asynchronously. Send POST \/Run\/{owner}\/{project} to start a task. Then poll POST \/Task\/Detail with a taskid until status reaches task_postprocess_end. Completed tasks return outputs with a contenttype and a url.<\/code><\/pre>\n<p><strong>Seed-V2 Mini<\/strong><\/p>\n<pre><code>-  Wiro executes models in an asynchronous manner -  To initiate a task, send a POST request to the `\/Run\/{owner}\/{project}` endpoint -  After starting a task, poll the `POST \/Task\/Detail` endpoint using a task ID until the task status reaches `task_postprocess_end` -  Once a task is completed, it will return output data -  Completed task outputs include a content type and a downloadable URL<\/code><\/pre>\n<p><strong>Qwen3.5-27B<\/strong><\/p>\n<pre><code>* Wiro processes model runs asynchronously.\n* Start a new task via POST \/Run\/{owner}\/{project}.\n* Poll POST \/Task\/Detail with the taskid to check status.\n* Stop polling when status reaches task_postprocess_end.\n* Completed tasks provide outputs containing a contenttype and url.<\/code><\/pre>\n<h3>Test 5: Time math<\/h3>\n<p><strong>Prompt<\/strong><\/p>\n<pre><code>A meeting starts at 09:15 and lasts 35 minutes. There is a 10 minute break. Then a second meeting lasts 50 minutes. What time does it end? Return only the time in HH:MM.<\/code><\/pre>\n<p><strong>Seed-V2 Mini<\/strong><\/p>\n<pre><code>10:50<\/code><\/pre>\n<p><strong>Qwen3.5-27B<\/strong><\/p>\n<pre><code>10:50<\/code><\/pre>\n<h2>When to pick which<\/h2>\n<ul>\n<li>Pick Seed-V2 Mini when a direct, clean response matters (JSON, snippets, short utilities).<\/li>\n<li>Pick Qwen3.5-27B when a task benefits from deeper internal reasoning, but make sure the client disables thinking output if the app needs strict formats.<\/li>\n<\/ul>\n<h2>Try them<\/h2>\n<p><a href=\"https:\/\/wiro.ai\/models\/bytedance\/seed-v2-mini\">https:\/\/wiro.ai\/models\/bytedance\/seed-v2-mini<\/a><br \/>\n<a href=\"https:\/\/wiro.ai\/models\/qwen\/qwen3-5-27b\">https:\/\/wiro.ai\/models\/qwen\/qwen3-5-27b<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"<p>Seed-V2 Mini vs Qwen3.5-27B sounds like a simple comparison. The outputs can look very different in practice. This post runs five small&hellip;<\/p>\n","protected":false},"author":4,"featured_media":1782,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[51],"tags":[88,70,100,148],"class_list":["post-1783","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-model-comparison","tag-bytedance","tag-llm","tag-qwen","tag-seed-v2"],"_links":{"self":[{"href":"https:\/\/wiro.ai\/blog\/wp-json\/wp\/v2\/posts\/1783","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/wiro.ai\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/wiro.ai\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/wiro.ai\/blog\/wp-json\/wp\/v2\/users\/4"}],"replies":[{"embeddable":true,"href":"https:\/\/wiro.ai\/blog\/wp-json\/wp\/v2\/comments?post=1783"}],"version-history":[{"count":1,"href":"https:\/\/wiro.ai\/blog\/wp-json\/wp\/v2\/posts\/1783\/revisions"}],"predecessor-version":[{"id":1799,"href":"https:\/\/wiro.ai\/blog\/wp-json\/wp\/v2\/posts\/1783\/revisions\/1799"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/wiro.ai\/blog\/wp-json\/wp\/v2\/media\/1782"}],"wp:attachment":[{"href":"https:\/\/wiro.ai\/blog\/wp-json\/wp\/v2\/media?parent=1783"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/wiro.ai\/blog\/wp-json\/wp\/v2\/categories?post=1783"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/wiro.ai\/blog\/wp-json\/wp\/v2\/tags?post=1783"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}