Codesota · Multimodal · Visual Question Answering · GQATasks/Multimodal/Visual Question Answering
Visual Question Answering · benchmark dataset · 2019 · EN

GQA: Visual Reasoning in the Real World.

22M compositional questions grounded in real images via scene graphs. Tests multi-step visual reasoning, spatial understanding, and attribute comparison.

Paper ↗Submit a result ↵
§ 01 · Leaderboard

Best published scores.

4 results indexed across 1 metric. Shaded row marks current SOTA; ties broken by submission date.

Primary
accuracy · higher is better
accuracy· primary
4 rows
#ModelOrgSubmittedPaper / codeaccuracy
01AIMv2 ViT-3B/14 + Llama 3.0 8B—Nov 2024Multimodal Autoregressive Pre-training of Large Vision E… · code73.30
02VideoLLaMA3 7B—Jan 2025VideoLLaMA 3: Frontier Multimodal Foundation Models for … · code64.90
03VideoLLaMA3 2B—Jan 2025VideoLLaMA 3: Frontier Multimodal Foundation Models for … · code62.70
04BLIP-2 ViT-g FlanT5 XXL—Jan 2023BLIP-2: Bootstrapping Language-Image Pre-training with F… · code44.70
Fig 2 · Rows sorted by score within each metric. Shaded row marks SOTA. Dates reflect model or paper release where available, otherwise the date Codesota accessed the source.
§ 03 · Progress

2 steps
of state of the art.

Each row below marks a model that broke the previous record on accuracy. Intermediate submissions are kept in the leaderboard above; only SOTA-setting entries are re-listed here.

Higher scores win. Each subsequent entry improved upon the previous best.

SOTA line · accuracy
  1. Jan 30, 2023BLIP-2 ViT-g FlanT5 XXL44.70
  2. Nov 21, 2024AIMv2 ViT-3B/14 + Llama 3.0 8B73.30
Fig 3 · SOTA-setting models only. 2 entries span Jan 2023 → Nov 2024.
§ 04 · Literature

3 papers
tied to this benchmark.

Every paper below corresponds to at least one row in the leaderboard above. Click through for the arXiv preprint and, when available, the reference implementation.

§ 06 · Contribute

Have a score that beats
this table?

Submit a checkpoint and a reproduction script. We will run it, publish the score, and — if it takes the top — annotate the step on the progress chart with your name.

Submit a result ↵Read submission guide
What a submission needs
  • 01A public checkpoint or API endpoint
  • 02A reproduction script with frozen commit + seed
  • 03Declared evaluation environment (Python, deps)
  • 04One row per metric declared by this dataset
  • 05A contact so we can follow up on discrepancies