chat-qwen3-6-35b-a3b
GitHub调用Qwen3.6-35B模型处理16K上下文内的文本内容提取及Agent工作流，支持结构化输出与工具调用。适用于普通多语言文本任务，但成本较高，建议优先使用chat-qwen。
Trigger Scenarios
Install
npx skills add zerogpu/zerogpu-router --skill chat-qwen3-6-35b-a3b -g -y
SKILL.md
Frontmatter
{
"name": "chat-qwen3-6-35b-a3b",
"description": "Chat with qwen3.6-35b-a3b, a 35B open-weight Mixture-of-Experts model with structured output and tool calling over a 16K-token context. Use for content extraction and agent workflows over a passage that fits in 16K tokens. Prefer chat-qwen for ordinary multilingual text — it is half the price on input and holds twice the context.",
"allowed-tools": "Bash(zerogpu chat_completions *)",
"argument-hint": "<text>"
}
Call qwen3.6-35b-a3b. $ARGUMENTS is the raw prompt — pass it verbatim, no escaping or quoting required (the heredoc below handles every shell metacharacter, newline, quote, and paren safely):
zerogpu chat_completions -m qwen3.6-35b-a3b <<'ZGPU_END_OF_INPUT'
$ARGUMENTS
ZGPU_END_OF_INPUT
At $0.20 / $1.50 per 1M input/output tokens this costs twice /zerogpu-router:chat-qwen on input and over three times on output, over half its context, so chat-qwen is the better default for text. Its 16,384-token window is the smallest of the chat models, so send the passage you want worked on rather than a whole document.
The model also accepts images and video, but this skill sends the request as text through chat_completions, so image and video prompts are out of its reach. Do not route them here.
Output is the assistant's answer as plain text — the model's reasoning trace comes back in a separate field and is not printed. Relay the answer as-is — do not rewrite or expand it.
Savings note: only if the command output literally contains a line starting with 💰 ZeroGPU savings, append that exact line, unchanged, as the last line of your reply. If no such line is present, say nothing about savings and do not mention or suggest /zerogpu-router:cost-savings — this note is intentionally occasional, not shown every time.
Version History
- 605eceb Current 2026-11-02 19:12
