AdaL
    Available now in AdaL CLI and Desktop

    Code with Qwen3.8 Flash
    in AdaL

    Alibaba's flash-tier multimodal MoE model and an early preview of the Qwen4 architecture. Use qwen3.8-flash in AdaL for fast, low-cost coding backed by a 262K native context window (extensible to 1M with YaRN).

    Get Started

    Run /model and pick Qwen3.8 Flash · $0.16/M input · $0.47/M output

    Why use Qwen3.8 Flash in AdaL

    Key capabilities and pricing

    Alibaba flash-tier · multimodal MoE

    125B total parameters with just 6B activated per token — unmatched cost-efficiency, per Qwen's announcement.

    Early Qwen4 architecture preview

    GDN + QSA hybrid attention, gated residual, N-gram embedding, and the Muon optimizer — a precursor to Qwen4.

    262K native context · 1M with YaRN

    Native 262K-token context window, extensible to 1M tokens via YaRN scaling.

    131K max output

    Same output ceiling as Qwen3.8 Max for substantial refactors.

    Thinking model

    Reasoning-capable for multi-step planning, coding, and analysis.

    Vision support

    Processes images and visual context alongside code.

    $0.16/M input · $0.47/M output

    Cached input at $0.016/M reduces repeat-context cost.

    How to pick Qwen3.8 Flash

    Available in AdaL CLI and Desktop through the model selector

    1

    Open AdaL CLI or AdaL Desktop

    Use the coding environment where AdaL already understands your repo, tools, and task context.

    2

    Run /model

    Open the model selector and choose your model from the list.

    3

    Start your task

    Ask AdaL to inspect, plan, edit, test, and review with your chosen model powering the agent loop.

    What to know

    Key details about Qwen3.8 Flash

    "Qwen3.8 Flash is a multimodal MoE model and an early preview of the Qwen4 architecture — open-weight, trained at 1/9 the cost of Qwen3.7-Plus while outperforming it, especially in coding and office tasks."

    Qwen

    "Per Qwen's announcement: 58.7 on DeepSWE 1.1, 62.5 on SWE-bench Pro, 73.9 on CoWorkBench, 84.5 on AndroidWorld, and 95.7 on MathVision (with CI)."

    Benchmarks

    "Use Qwen3.8 Flash for fast iteration where a large native context window and vision support matter but budget is tight."

    AdaL

    "262K native context, extensible to 1M tokens with YaRN."

    Context

    "Available now in AdaL CLI and Desktop, 50% off launch week."

    Availability

    50% off launch week

    Ready to try Qwen3.8 Flash in AdaL?

    Open AdaL CLI or Desktop, run /model, and choose Qwen3.8 Flash.

    Get Started with AdaL
    AdaL CLIAdaL Desktop