QwenSpecialist modelOpen weights
Qwen-Drive-1.0-4B
42out of 100same as last month
Summary
Qwen-Drive-1.0-4B is an open vision-language model for autonomous driving. A Qwen3.5-4B VLM remains the shared core, while a BEV perception head adds 3D detection, semantic occupancy, and map segmentation. A Planning Expert generates future vehicle trajectories.
The Hugging Face release includes the VLM checkpoint plus separate planner-sft, planner-rl, and perception modules. The Apache-2.0 weights run locally. The model is specialized for driving scenes, perception, visual question answering, and motion planning rather than general chat tasks.
All benchmark results
Every published result with its source, test setup and comparison models.
Show all 114 published benchmark results
Page 1 of 10
Measurements without matching peer values
These measurements have no matching peer values under the same test conditions. Their original values and sources remain available here.
NAVSIM no-collision score (Show measurement, test conditions, and source)
- Source value
- 98.4
- Score
- 98.4
- Metric
- NC, planner SFT with reasoning
- Unit
- %
- Benchmark version
- NAVSIM v1.1 navtest
- Category
- autonomous-driving
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Qwen-Drive-1.0 official evaluation report
- Status
- active
- Retrieved at
- 2026-10-16
- Methodology
- Vendor-reported result. Qwen-Drive-1.0 planner SFT with reasoning; NAVSIM v1.1 navtest component metric.
NAVSIM drivable-area compliance (Show measurement, test conditions, and source)
- Source value
- 96.6
- Score
- 96.6
- Metric
- DAC, planner SFT with reasoning
- Unit
- %
- Benchmark version
- NAVSIM v1.1 navtest
- Category
- autonomous-driving
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Qwen-Drive-1.0 official evaluation report
- Status
- active
- Retrieved at
- 2026-10-16
- Methodology
- Vendor-reported result. Qwen-Drive-1.0 planner SFT with reasoning; NAVSIM v1.1 navtest component metric.
NAVSIM ego progress (Show measurement, test conditions, and source)
- Source value
- 82.4
- Score
- 82.4
- Metric
- EP, planner SFT with reasoning
- Unit
- %
- Benchmark version
- NAVSIM v1.1 navtest
- Category
- autonomous-driving
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Qwen-Drive-1.0 official evaluation report
- Status
- active
- Retrieved at
- 2026-10-16
- Methodology
- Vendor-reported result. Qwen-Drive-1.0 planner SFT with reasoning; NAVSIM v1.1 navtest component metric.
NAVSIM time-to-collision score (Show measurement, test conditions, and source)
- Source value
- 94.7
- Score
- 94.7
- Metric
- TTC, planner SFT with reasoning
- Unit
- %
- Benchmark version
- NAVSIM v1.1 navtest
- Category
- autonomous-driving
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Qwen-Drive-1.0 official evaluation report
- Status
- active
- Retrieved at
- 2026-10-16
- Methodology
- Vendor-reported result. Qwen-Drive-1.0 planner SFT with reasoning; NAVSIM v1.1 navtest component metric.
NAVSIM comfort score (Show measurement, test conditions, and source)
- Source value
- 100
- Score
- 100
- Metric
- Comfort, planner SFT with reasoning
- Unit
- %
- Benchmark version
- NAVSIM v1.1 navtest
- Category
- autonomous-driving
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Qwen-Drive-1.0 official evaluation report
- Status
- active
- Retrieved at
- 2026-10-16
- Methodology
- Vendor-reported result. Qwen-Drive-1.0 planner SFT with reasoning; NAVSIM v1.1 navtest component metric.
NAVSIM no-collision score (Show measurement, test conditions, and source)
- Source value
- 98.6
- Score
- 98.6
- Metric
- NC, planner RL
- Unit
- %
- Benchmark version
- NAVSIM v1.1 navtest
- Category
- autonomous-driving
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Qwen-Drive-1.0 official evaluation report
- Status
- active
- Retrieved at
- 2026-10-16
- Methodology
- Vendor-reported result. Qwen-Drive-1.0 planner RL; NAVSIM v1.1 navtest component metric.
NAVSIM drivable-area compliance (Show measurement, test conditions, and source)
- Source value
- 98.2
- Score
- 98.2
- Metric
- DAC, planner RL
- Unit
- %
- Benchmark version
- NAVSIM v1.1 navtest
- Category
- autonomous-driving
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Qwen-Drive-1.0 official evaluation report
- Status
- active
- Retrieved at
- 2026-10-16
- Methodology
- Vendor-reported result. Qwen-Drive-1.0 planner RL; NAVSIM v1.1 navtest component metric.
NAVSIM ego progress (Show measurement, test conditions, and source)
- Source value
- 84.8
- Score
- 84.8
- Metric
- EP, planner RL
- Unit
- %
- Benchmark version
- NAVSIM v1.1 navtest
- Category
- autonomous-driving
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Qwen-Drive-1.0 official evaluation report
- Status
- active
- Retrieved at
- 2026-10-16
- Methodology
- Vendor-reported result. Qwen-Drive-1.0 planner RL; NAVSIM v1.1 navtest component metric.
NAVSIM time-to-collision score (Show measurement, test conditions, and source)
- Source value
- 95.9
- Score
- 95.9
- Metric
- TTC, planner RL
- Unit
- %
- Benchmark version
- NAVSIM v1.1 navtest
- Category
- autonomous-driving
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Qwen-Drive-1.0 official evaluation report
- Status
- active
- Retrieved at
- 2026-10-16
- Methodology
- Vendor-reported result. Qwen-Drive-1.0 planner RL; NAVSIM v1.1 navtest component metric.
NAVSIM comfort score (Show measurement, test conditions, and source)
- Source value
- 100
- Score
- 100
- Metric
- Comfort, planner RL
- Unit
- %
- Benchmark version
- NAVSIM v1.1 navtest
- Category
- autonomous-driving
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Qwen-Drive-1.0 official evaluation report
- Status
- active
- Retrieved at
- 2026-10-16
- Methodology
- Vendor-reported result. Qwen-Drive-1.0 planner RL; NAVSIM v1.1 navtest component metric.
AlpaSim close encounters (Show measurement, test conditions, and source)
- Source value
- 38
- Score
- 38
- Metric
- all close encounters; planner SFT with reasoning; lower is better
- Unit
- %
- Benchmark version
- PAI-AV-NuRec v26.02
- Category
- autonomous-driving
- Direction
- Lower is better
- Sample
- 916
- Source type
- vendor-reported
- Evaluator
- Qwen-Drive-1.0 official evaluation report
- Status
- active
- Retrieved at
- 2026-10-16
- Methodology
- Vendor-reported result. AlpaSim, PAI-AV-NuRec v26.02; 916 scenarios; planner SFT with reasoning.
AlpaSim at-fault close encounters (Show measurement, test conditions, and source)
- Source value
- 12
- Score
- 12
- Metric
- at-fault close encounters; planner SFT with reasoning; lower is better
- Unit
- %
- Benchmark version
- PAI-AV-NuRec v26.02
- Category
- autonomous-driving
- Direction
- Lower is better
- Sample
- 916
- Source type
- vendor-reported
- Evaluator
- Qwen-Drive-1.0 official evaluation report
- Status
- active
- Retrieved at
- 2026-10-16
- Methodology
- Vendor-reported result. AlpaSim, PAI-AV-NuRec v26.02; 916 scenarios; planner SFT with reasoning.
Page 1 of 10
Popularity compared
How often the model is searched for, used and mentioned, compared with every model of its category.
Popularity index
0 to 100, higher is more popular
20 of 29 models
As of November 2026 · Source: gradually.ai popularity index
Over time
Weekly values from October 18, 2026 to November 5, 2026
Qwen-Drive-1.0-4BAstaBrief-8BTeleOCRpplx-embed-v2-context-9b-previewAgora-2
All specifications
Architecture
Watermarking and provenance
AI watermarking refers to invisible signals embedded in the output. Visible marks and provenance metadata are listed separately. Unknown means there is insufficient evidence for that particular field.
- AI watermarking
- Unknown
- no source
- Watermarking technique
- Not known
- no source
Licence and access
API
Further details
Sources and data date
Show all 14 sources
Benchmark sources at a glance
Which independent evaluators tested this model. This is coverage, not a quality ranking.
No active result from the reviewed external evaluators can currently be attributed to this model. Other source evidence remains available below.
Without vendor claims and historical or disputed results; those stay labelled in the single values.
| Type | Evidence and data date |
|---|---|
| Research status | Research date October 4, 2026. 7 source URLs checked. This documents the inspected sources, not an exhaustive inventory of every publication. |
| Unresolved | checkpoint-component-comparability Note The published repository-root Qwen3.5-4B VLM serves VQA by itself, and the VQA/general-VLM rows below are model-only evaluations of that component. The planner SFT/RL scores require their respective sibling Planning Experts; perception scores require the sibling BEV head. Those system/component results are not interchangeable with the root-VLM VQA scores or directly comparable as bare model-only scores. |
| Not published | row-level-test-populations Note Individual test populations are not reported for NAVSIM or Waymo table rows. The PhysicalAI 644-example split and AlpaSim 916-scenario set are retained only where the report states them. |
| Not published | driving-vqa-row-populations Note The paper reports exact Driving-VQA scores and identifies its judge protocols, but does not publish per-row sample counts for these evaluations. |
| Not published | language2drive-confidence-interval Original wording The paper reports point values and test-unit counts but no confidence intervals for these rows. |
| Additional source | Qwen-Drive model card (retrieved October 3, 2026; October 16, 2026) · Qwen-Drive model card · Editorial description reviewed October 3, 2026 |
| Additional source | Qwen-Drive announcement (retrieved October 3, 2026) · Qwen-Drive announcement · Editorial description reviewed October 3, 2026 |
| Additional source | Qwen-Drive technical report (retrieved October 3, 2026; October 15, 2026) · Qwen-Drive technical report · Editorial description reviewed October 3, 2026 |
| Additional source | Qwen-Drive-1.0 official evaluation report (retrieved October 16, 2026; October 19, 2026) |
| Additional source | Qwen-Drive-1.0 technical report (retrieved October 16, 2026) |
| Additional source | Language2Drive paper (retrieved October 17, 2026) |
| Additional source | Qwen-Drive-1.0-4B model card, motion-planning table (retrieved October 23, 2026) |
| Additional source | Model metadata (retrieved October 15, 2026) |
| Additional source | OpenRouter: Popularity index (derived), CC BY 4.0 |
| Additional source | Vercel AI Gateway: Popularity index (derived), CC BY 4.0 |
| Additional source | Ollama: cumulative downloads for the popularity index |
| Additional source | Replicate: cumulative runs for the popularity index |
| Additional source | Google Trending Now: limited feed selection for Germany and the US |
| Additional source | Creative Commons Attribution 4.0 |