Qwen-Drive 1.0
Qwen-Drive1.0 is a vision-language foundation model specialized for autonomous-driving research while retaining free-form visual question answering and instruction following. Its Qwen3.5-4B language/vision backbone is paired with an external bird’s-eye-view perception head and a flow-matching Planning Expert for future trajectories. General vision-language supervision is mixed with driving data to preserve broad visual understanding. The August31 technical report precedes the public weight rollout. Apache-2.0 weights and demo data support research; SFT and reward-optimized planners are grouped as variants.
2026-08-31
4B VLM backbone, plus external perception/planning modules
Qwen3.5 hybrid-attention vision-language backbone with external BEV perception and flow-matching planner
Apache-2.0
Specifications
- Parameters
- 4B VLM backbone, plus external perception/planning modules
- Architecture
- Qwen3.5 hybrid-attention vision-language backbone with external BEV perception and flow-matching planner
- License
- Apache-2.0
- Type
- multimodal
- Modalities
- textimage
Benchmark Scores
Advanced Specifications
- Model Family
- Qwen
- Finetuned From
- Qwen3.5-4B
- API Access
- Not Available
- Chat Interface
- Not Available
- Variants
- planner-sftplanner-rl
Capabilities & Limitations
- Capabilities
- visual question answeringdriving-scene reasoning3D perceptionmotion planning
- Known Limitations
- A research checkpoint is not a validated autonomous-driving control system.Planning and perception require the provided task-specific heads and data preprocessing.
- Notable Use Cases
- driving-model researchscene-question answeringmultimodal planning experiments