← Back to Docs

Qwen 3 30B

Alibaba's general-purpose reasoning model with strong multi-turn capabilities.

ProviderAlibaba Cloud (Qwen team)
Parameters30 billion
Context128K tokens
Best forReasoning, planning, multi-step analysis
Dispatch rolespec_agent, verify_agent
InfrastructureDispatch Pro cloud
CostIncluded in Pro ($20/month, shared 5M token pool)

Overview

Qwen 3 30B is a general-purpose reasoning model that balances capability with efficiency. Unlike the code-specialized Qwen Coder, Qwen 3 excels at higher-level tasks: planning implementation approaches, generating test strategies, and reasoning about architectural decisions. In Dispatch, it's a versatile choice for any pipeline stage that needs careful thinking.

When Dispatch uses Qwen 3

  • Spec generation -- Architectural planning and goal decomposition.
  • Test spec generation -- Comprehensive test strategy with coverage analysis.
  • Verification reasoning -- Multi-step correctness analysis.

Configuration

Available to all Dispatch Pro subscribers. Set any pipeline stage to qwen-3-30b in Settings > Pipeline Profiles, or let Auto Model select based on task requirements.

Isolation

All cloud model runs use worktree isolation.