Qwen 3.6 Models Are Now Available in Puter.js

Puter.js now supports the full Qwen 3.6 model series from Alibaba's Qwen team — joining the previously released Qwen 3.6 Plus with a new proprietary flagship, two open-weight models, and a speed-optimized Flash tier.

What is Qwen 3.6?

Qwen 3.6 is Alibaba's next-generation model family, purpose-built for agentic coding and repository-level reasoning. It builds on a hybrid architecture combining linear attention with sparse mixture-of-experts routing, and introduces Thinking Preservation — a feature that retains reasoning traces across conversation turns to reduce redundant computation in multi-step agent loops.

• Qwen3.6 Max Preview: The proprietary flagship released on April 20, 2026. Claims the top score on six major coding benchmarks including SWE-bench Pro, Terminal-Bench 2.0, SkillsBench, and SciCode. Supports a 256K context window with an API compatible with both OpenAI and Anthropic specs.
• Qwen3.6 Plus: A 1M-token-context flagship optimized for agentic coding and complex multi-step workflows. Scores 78.8 on SWE-bench Verified and leads MCPMark at 48.2%.
• Qwen3.6-27B: A dense 27B-parameter open-weight multimodal model. Scores 77.2% on SWE-bench Verified and 59.3% on Terminal-Bench 2.0, outperforming the previous-gen Qwen3.5-397B-A17B despite being far smaller. Released under Apache 2.0.
• Qwen3.6-35B-A3B: A sparse MoE with 35B total and just 3B active parameters. Scores 73.4% on SWE-bench Verified and 51.5% on Terminal-Bench 2.0 — frontier-adjacent coding performance at a fraction of the compute. Open-weight under Apache 2.0.
• Qwen3.6 Flash: The speed-optimized tier, designed for high-throughput, low-latency workloads like classification, extraction, and summarization.

Examples

Agentic coding with the new flagship

puter.ai.chat("Refactor this React form to use react-hook-form with proper validation, error states, and accessibility attributes",
  { model: 'qwen/qwen3.6-max-preview' }
);

Repository-level reasoning on open weights

puter.ai.chat("Trace how a request flows from the API gateway through the auth middleware to the database layer in this codebase, and identify any places state is mutated unsafely",
  { model: 'qwen/qwen3.6-27b' }
);

Efficient inference with sparse MoE

puter.ai.chat("Write a Python script that ingests a directory of PDFs, extracts tables, and writes them to a SQLite database",
  { model: 'qwen/qwen3.6-35b-a3b' }
);

High-throughput summarization

puter.ai.chat("Summarize this customer support ticket in two sentences",
  { model: 'qwen/qwen3.6-flash' }
);

Streaming with thinking preservation

const response = await puter.ai.chat(
  "Design a rate limiter that handles 100K requests/sec across multiple regions, walk through the trade-offs",
  { model: 'qwen/qwen3.6-max-preview', stream: true }
);

for await (const part of response) {
  if (part?.reasoning) puter.print(part?.reasoning);
  else puter.print(part?.text);
}

Get Started Now

// npm install @heyputer/puter.js
import { puter } from '@heyputer/puter.js';

Or add one script tag to your HTML:

<script src="https://js.puter.com/v2/"></script>

No API keys needed. Start building with Qwen 3.6 models immediately.

Learn more:

• Free, Unlimited Qwen API
• Puter.js AI Documentation