Back to reviews
Qwen3.6-Plus

Qwen3.6-Plus

The agentic coding model beating Claude Opus 4.5 — free on OpenRouter

Qwen3.6-Plus is Alibaba's latest frontier model, built specifically for agentic real-world tasks with a particular emphasis on software engineering. Released in preview on OpenRouter as a free tier, it scores 61.6 on Terminal-Bench 2.0, edging past Claude Opus 4.5 (59.3), while running at roughly 3x the speed. It supports a 1M token context window with 65K output tokens — larger than most competitors. Under the hood, Qwen3.6-Plus is a sparse mixture-of-experts architecture, activating a fraction of its parameters per forward pass for efficiency. It supports both text and multimodal inputs, and the API supports tool use natively — making it well-suited for agent loops. The free preview is positioned as a direct challenge to OpenAI and Anthropic in the agentic coding space. The timing is notable: released the same week as Google Gemma 4 and Cursor 3, signaling an industry-wide pivot from autocomplete to full autonomous agents. With free preview access already expiring, Alibaba is clearly using the buzz from benchmark dominance to drive early adoption at the API tier.

Panel Reviews

The Builder

The Builder

Developer Perspective

Ship

The Terminal-Bench numbers don't lie — this thing completes agentic coding tasks better than Opus at a fraction of the cost. The 1M context window means I can throw an entire monorepo at it. Free preview while it lasts is a no-brainer for any dev working on agent pipelines.

The Skeptic

The Skeptic

Reality Check

Skip

Benchmark performance on Terminal-Bench doesn't always translate to real-world reliability. Alibaba's track record on model longevity and API uptime is spottier than Anthropic's or OpenAI's. The free preview ending today is also a classic bait-and-switch move — the real question is what the paid tier costs.

The Futurist

The Futurist

Big Picture

Ship

We're seeing the first real multi-model agent race, and Qwen3.6-Plus is the opening shot from China. The combination of 1M context, agentic optimization, and benchmark-beating performance signals that the era of Western AI dominance in coding agents may be over. This reshapes the market.

The Creator

The Creator

Content & Design

Ship

For automation-heavy creative workflows — building tools, scraping, image pipelines — having a faster, cheaper frontier model with giant context is genuinely useful. I can run whole project contexts through it without hitting limits. The free preview makes it a zero-cost experiment.

Community Sentiment

Overall1,047 mentions
67% positive23% neutral10% negative
Hacker News187 mentions

Terminal-Bench scores vs Opus 4.5

Reddit340 mentions

1M context window and free OpenRouter access

Twitter/X520 mentions

Free preview availability and benchmark claims