Benchmark Almanac

MotoVerse

SlowFast 51.1% tops its leaderboard Qwen3-VL-2B-Instruct 46.3% leads the models we track
Not in model cards

Models use a rider's camera to anticipate maneuvers and predict nearby road users' trajectories. The authors compare vision, language, and conventional prediction models.

Released
November 2026
Kind
Embodied reasoning
Level
Everyday
Scored
Checked automatically
Answers
Prediction
Tasks
All public
Measures
Capability
Language
English
In and out
Video → Time series, Text
Interaction
Single turn
Items from
Real work and use
Refresh
One fixed set
Run by
Anyone
In views
Vision
Last model added
October 2025
Scores refreshed
November 20th, 2026
12
Entries in its paper, 4 models we track

SlowFast tops the leaderboard with 51.1% Qwen3-VL-2B-Instruct leads with 46.3% among models released by October 2025

Best score per entry from the benchmark's paper, read on November 20th, 2026. A paper’s table doesn’t add newer models. The headline is maneuver anticipation accuracy after observing the first 20% of the maneuver. Where a model ran with several settings or harnesses, we keep its best. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. The best model we track is Qwen3-VL-2B-Instruct, at 46.3%. This view shows only models we track, each at its best.

The newest model with a score here came out in October 2025, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:

  1. 1 SlowFast Fine-tuned 51.1%
  2. 2 UniDriveVLA LoRA 50.4%
  3. 3 X3D Fine-tuned 49.3%
  4. 4 I3D Fine-tuned 48.5%
  5. 5 C2D Fine-tuned 48.1%
  6. 6 Qwen3-VL-2B-Instruct LoRA 46.3%
  7. 7 InternVL2.5-1B Zero-shot 31.5%
  8. 8 InternVL2.5-4B Zero-shot 30%
  9. 9 Cosmos-Reason2-8B Zero-shot 21.9%
  10. 10 DriveLMM-o1 Zero-shot 21.5%
  11. 11 Qwen3-VL-4B-Instruct Zero-shot 17%
  12. 12 Alpamayo-1.5-10B Zero-shot 9.00%

Source: The MotoVerse paper

#EntryScoreSettingDateSource
1 SlowFast 51.1% Fine-tuned Official leaderboard
2 UniDriveVLA 50.4% LoRA Official leaderboard
3 X3D 49.3% Fine-tuned Official leaderboard
4 I3D 48.5% Fine-tuned Official leaderboard
5 C2D 48.1% Fine-tuned Official leaderboard
6 Qwen3-VL-2B-Instruct 46.3% LoRA Oct 2025 Official leaderboard
7 InternVL2.5-1B 31.5% Zero-shot Dec 2024 Official leaderboard
8 InternVL2.5-4B 30% Zero-shot Dec 2024 Official leaderboard
9 Cosmos-Reason2-8B 21.9% Zero-shot Official leaderboard
10 DriveLMM-o1 21.5% Zero-shot Official leaderboard
11 Qwen3-VL-4B-Instruct 17% Zero-shot Oct 2025 Official leaderboard
12 Alpamayo-1.5-10B 9.00% Zero-shot Official leaderboard
#ModelScoreSettingDateSource
1 Qwen3-VL-2B-Instruct 46.3% LoRA Oct 2025 Official leaderboard
2 InternVL2.5-1B 31.5% Zero-shot Dec 2024 Official leaderboard
3 InternVL2.5-4B 30% Zero-shot Dec 2024 Official leaderboard
4 Qwen3-VL-4B-Instruct 17% Zero-shot Oct 2025 Official leaderboard

Rankings by part

We read 18 parts from the benchmark’s paper in 5 groups: 20% observation, 30% observation, 50% observation, 100% observation and trajectory prediction. Each ranks models on one part of the test. Qwen3-VL-2B-Instruct leads every one of them. Arrows show how far a model moved from its overall rank among the models we track.

Accuracy

4 models we track have a score here.

  1. 1 Qwen3-VL-2B-Instruct 46.3%
  2. 2 InternVL2.5-1B 31.5%
  3. 3 InternVL2.5-4B 30%
  4. 4 Qwen3-VL-4B-Instruct 17%

Number one over time

Lab claims count from the day the card came out, independent runs from the model’s release.

Qwen
Benchmark paper 1 leader

Source: The MotoVerse paper

Scores over time

All the scores we have, one dot each, so a model tested at several settings appears more than once.

Plot with other benchmarks
In the benchmark’s paper Best so far
0% 20% 40% 60% 80% 100% 2025 2026 Near the ceiling (85%) Qwen3-VL-2B-Instruct InternVL2.5-1B 0% 20% 40% 60% 80% 100% 2025 2026 Near the ceiling (85%) Qwen3-VL-2B-Instruct InternVL2.5-1B

Source: Official MotoVerse leaderboard

No lab model card reports it yet

We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.

More in Robotics

All 172 in Robotics
  • ERQA In use 400 multiple-choice visual questions from robot and egocentric footage on spatial, trajectory, action and state reasoning. Accuracy. Gemini Robotics ER 2 78.5%
  • SimplerEnv In use Real-to-sim replicas of Google Robot and WidowX/Bridge setups. Policies scored by success rate on pick, move, drawer and Bridge tasks. Praxis-1 80%
  • RoboLab In use Zero-shot DROID Franka policies in Isaac Lab on 120 tabletop tasks at three instruction levels, 10 episodes each. Success rate. Praxis-1 46.2%
  • RoboDojo In use Generalist manipulation policies run 42 Isaac Sim tasks and 18 remote real-robot tasks. Scored by task score and success rate. GPT-6 Astra 27.6%
  • Progress classification (Google DeepMind) In use Models assign each frame of a robot video to one of five task-progress bands from 0 to 100 percent. Scored by accuracy. Gemini Robotics ER 2 57.4%
  • Human proximity (Google DeepMind robotics) In use Models judge from robot camera views whether a human is within 1 meter, a spatial awareness check for safe robot behavior. Scored by accuracy. Gemini Robotics ER 2 93%

Data audit

9 of 10 checks pass, 1 warning. Look first at the Newest score check. The newest came 13 months ago, in October 2025, for Qwen3-VL-2B-Instruct. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 1 warning 3 pass 1 note
  • Pass Independent runs Scores from the benchmark’s paper (12 entries).
  • Pass Models scored 4 models we track, plus 8 other entries on its leaderboard.
  • Warning Newest score The newest came 13 months ago, in October 2025, for Qwen3-VL-2B-Instruct.
  • Note Frontier models None of its models are on the capabilities index.
  • Pass Headroom The best is 46.3%, under 85%, so it still separates models.
Leaderboard import 3 pass
  • Pass Importer We read the board on arxiv.org: 5 rows for models we track and 8 other entries at the last read.
  • Pass Last read November 20th, 2026.
  • Pass Read worked The last read worked.
Consistency 1 pass 1 note
  • Note Lab claims No model has both a lab number and an independent run.
  • Pass Known issues No published audit found problems.

How we check