Currently listed through these providers:
Model details
Qwen3.6 Max Preview
Qwen3.6 Max Preview is an early-access release from Alibaba Cloud that builds on the line introduced with Qwen3.6-Plus, positioning itself as a stronger, knowledge-rich successor within the same family. According to Alibaba's community blog and third-party coverage, it is offered as a proprietary, closed-weights frontier model rather than an open release, and is being surfaced through multiple gateways and platforms during its preview window. This positioning suggests Alibaba is using the preview phase to gather feedback and demonstrate capability gains before a broader rollout, while still gating direct access behind API channels rather than downloadable weights.
Independent listings converge on the model's technical shape: a sparse mixture-of-experts design with roughly one trillion total parameters, paired with a 262,144-token context window and a 65,536-token maximum output. The Qubrid AI write-up reports that within days of launch the preview claimed the top spot on six major programming benchmarks and ranked second overall on the Artificial Analysis Intelligence Index, while also displacing the prior leader on the ToolcallFormatIFBench instruction-following leaderboard. Together, those results frame the preview as a reasoning- and coding-oriented model aimed at complex multi-step workflows, long-context analysis, and tool-driven agent tasks where stronger world knowledge and instruction-following are the practical selling points.
Quick Info
Powered by- Provider
- DevPass (LLM Gateway)
- Model key
- qwen3.6-max-preview
- Release date
- Apr 20, 2026
- Last updated
- Apr 20, 2026
- Knowledge cutoff
- 2025-04
- Input modalities
- Output modalities
- Capabilities
Cost
- Input token cost
- $1.30
- Output token cost
- $7.80
Limits
- Output tokens
- 65,536 tokens
- Context window
- 262,144 tokens