AWS's Strands Labs has released Strands Decider 2B, an open-source decision model designed to accelerate agentic AI workflows by handling option-selection tasks in approximately 115 milliseconds1,2.
The model was fine-tuned from an Alibaba Qwen base and is positioned as a free, open-source competitor to Jev, according to VentureBeat's Carl Franzen3. Strands Labs, described as AWS's experimental agent-development project, is making the model available to the open-source community.
A new model class
Strands Decider 2B belongs to an emerging class of lightweight AI systems that AWS calls "decision models". ANALYSIS Rather than generating long-form text answers, these models are built to pick among options quickly, a function that becomes a bottleneck when orchestrating multi-step agentic pipelines. The "2B" designation indicates a two-billion-parameter scale, placing it firmly in the category of models small enough to run on modest infrastructure.
The roughly 115-millisecond inference latency reported by MarkTechPost suggests the model is optimized for the rapid, repeated calls that agentic frameworks require when routing tasks, selecting tools, or choosing next steps. ◆ At that speed, the decision layer would add minimal overhead to an agent loop, a practical consideration for production deployments where each additional millisecond compounds across dozens of orchestration steps.
Competitive positioning
The VentureBeat report frames Strands Decider 2B as a direct competitor to Jev, noting that "the race to make AI agents cheaper and faster is moving beyond the models that write their answers". ◆ By open-sourcing the model and building on Alibaba's Qwen weights, AWS lowers the barrier for developers who want to integrate a purpose-built decision layer without licensing fees or vendor lock-in.
Strands Labs' choice of the Qwen base is notable: it pairs an AWS-branded agent toolchain with a foundation model originating from Alibaba's ecosystem. ◆ The cross-ecosystem fine-tuning reflects a pragmatic approach in which the selection of base weights is driven by fit-for-task rather than corporate affiliation.
The release lands during a week of large-scale capital activity across the AI industry. OpenAI is in talks for a $30 billion pre-IPO round at roughly $1.4 trillion valuation[2], and Anthropic's IPO timeline is drawing prediction-market attention[1].
Strands Decider 2B is available now as an open-source release.