Qwen 3.6 Plus
Alibaba's flagship closed Qwen model via the Vercel AI Gateway. 396B MoE with function calling and image input support. Available serverless on just4o.chat at the base tier. Uses 4 base requests per send before length multipliers. Does not support web search.
You always get the exact model you pick — we never silently route you to another.
About Qwen 3.6 Plus
At $0.50 per million input tokens, Qwen 3.6 Plus punches well above its price band — scoring 78.8 on SWE-bench Verified and 61.6 on Terminal-Bench 2.0, where it outpaces Claude 4.5 Opus on agentic coding tasks. The 1 million token context window lets you drop in entire codebases for security audits, multi-file refactors, or long-horizon agent sessions without chunking or worrying about cost. Always-on chain-of-thought reasoning is baked into the architecture rather than toggled per request, and native tool-calling makes it well-suited for multi-step workflows. Developers building high-volume API applications have reported generating hundreds of millions of tokens during its preview period — its first-day usage crossed one trillion tokens across platforms. That said, the long context is not a silver bullet: retrieval accuracy degrades in the middle of very long inputs, and real-world testing has surfaced instruction-following inconsistencies and occasional tool-calling failures that more mature providers handle more reliably. For cost-sensitive production deployments where coding and document analysis are the core workload, few models compete at this price.
Best for
- Agentic coding and repository-level problem solving — ingest full codebases for security audits, multi-file bug fixes, and automated refactoring
- Long document analysis — contract review, enterprise knowledge management, and processing dense technical documentation within a 1M token window
- Multi-step tool-calling workflows — orchestrate agent pipelines that require reliable native function execution across many steps
- Cost-sensitive production API deployments — high-volume applications that need solid performance without frontier-model pricing
- Multilingual applications — technical content generation and analysis across 119 supported languages
Specifications
| Provider | Alibaba |
|---|---|
| Released | 2026-03-31 |
| Intelligence | High |
| Speed | Medium |
| Context window | 1,000,000 tokens |
| Max output | 65,536 tokens |
| Knowledge cutoff | Not officially published |
| Input price | $0.50 / 1M tokens |
| Output price | $3.00 / 1M tokens |
| Request cost | 4 base requests |
| Plan tier | Base |
| Input | Text, Image |
| Output | Text |
| Features | Cached input: $0.10 / 1M tokens, 4 base requests per send before length multipliers, Function calling supported, Served through the Vercel AI Gateway (alibaba/qwen3.6-plus), Zero data retention and no prompt training |
| Model ID | qwen3p6-plus |
Frequently asked questions
Input is $0.50 per million tokens and output is $3.00 per million tokens. Cache hits drop to $0.05 per million tokens — a 90% discount — making repeated context reuse very economical for long agent sessions.
1 million tokens. The first 256K tokens use the model's native architecture; beyond that, a YaRN extension handles the remainder. Retrieval accuracy can degrade in middle portions of very long contexts, so plan accordingly.
Agentic coding tasks, repository-level code analysis, long document processing, and multi-step tool-calling workflows. Its 78.8 SWE-bench Verified score and 61.6 Terminal-Bench 2.0 score reflect genuine strength in software engineering.
Yes — users have reported instruction-following inconsistencies, occasional tool-calling failures, and hallucinations in some scenarios. Performance in the middle of very long contexts is weaker than at the start or end. Some testers also found it weaker than its predecessor Qwen 3.5 Plus on certain practical tasks.
Qwen 3.6 Plus improves on agentic coding benchmarks and adds a larger context window, but some users in real-world tests have found 3.5 Plus more consistent on general tasks. The right choice depends on whether you need the extended context and agentic capabilities.
No. Unlike earlier Qwen releases, 3.6 Plus is proprietary and API-only through Alibaba Cloud DashScope and OpenAI-compatible endpoints. Weights are not publicly available, which limits transparency and data residency options.