VTB leaderboard
As of 6 Nov 2026, the top score on VTB is 44.8% by Muse Spark 1.1 (Meta), ahead of GPT-5.4 at 29.2%, across 21 model configurations with a published result (Scale AI SEAL, result dated 9 Jul 2026). Practical vision tasks, from reading charts to spotting differences, graded against references. VTB is shown for reference only and does not count toward the BenchLeader Index.
Key facts (as of 6 Nov 2026)
- Muse Spark 1.1 holds the top VTB score at 44.8% (Scale AI SEAL, 9 Jul 2026).
- GPT-5.4 is second on VTB at 29.2% (Scale AI SEAL, 24 Mar 2026).
- VTB: Scale's vision task benchmark across practical image tasks. Scale AI (Scale AI SEAL).
- 21 model configurations have a published VTB result on BenchLeader (BenchLeader, 6 Nov 2026).
- VTB does not feed the BenchLeader Index; it is shown for reference (BenchLeader methodology).
Scale's vision task benchmark across practical image tasks. Scale AI.
- Published by
- Scale AI SEAL
- Category
- Multimodal
- Index weight
- Reference only
- Models
- 21
- Data as of
- 6 Nov 2026
Scale AI SEAL Leaderboards.
What the test looks like
Practical vision tasks, from reading charts to spotting differences, graded against references.
How it is scored
Accuracy, published by Scale AI.
What to keep in mind
Vision models only.
- 1Muse Spark 1.144.8%
- 2GPT-5.4 (high)29.2%
- 3Gemini 3.1 Pro29.0%
- 4Claude Opus 4.6 (thinking)27.5%
- 5Gemini 3 Pro26.9%
- 6GPT-5 (thinking)18.7%
- 7GPT-517.0%
- 8o313.7%
- 9Gemini 2.5 Pro11.8%
- 10o4-mini11.1%
- 11Claude Sonnet 4.5 (thinking)6.2%
- 12Claude Sonnet 4.55.6%
- 13GPT-4.15.5%
- 14Claude Opus 4.1 (thinking)5.2%
- 15Claude Opus 4.14.7%
| # | ||||
|---|---|---|---|---|
| 1 |  Muse Spark 1.1 | 44.8% | 62.2 | 2026-07-09 |
| 2 |  GPT-5.4high | 29.2% | 58.8 | 2026-03-24 |
| 3 |  Gemini 3.1 Pro | 29.0% | 60.5 | 2026-03-24 |
| 4 |  Claude Opus 4.6thinking | 27.5% | 62.4 | 2026-03-24 |
| 5 | Gemini 3 Pro | 26.9% | 60.2 | 2025-11-20 |
| 6 | GPT-5thinking | 18.7% | – | 2025-10-08 |
| 7 | GPT-5 | 17.0% | 57.7 | 2025-10-08 |
| 8 | o3 | 13.7% | 61.0 | 2025-10-08 |
| 9 | Gemini 2.5 Pro | 11.8% | 54.5 | 2025-10-08 |
| 10 | o4-mini | 11.1% | 56.9 | 2025-10-08 |
| 11 | Claude Sonnet 4.5thinking | 6.2% | 52.2 | 2025-10-08 |
| 12 | Claude Sonnet 4.5 | 5.6% | 53.7 | 2025-10-08 |
| 13 | GPT-4.1 | 5.5% | 48.5 | 2025-10-08 |
| 14 | Claude Opus 4.1thinking | 5.2% | 57.8 | 2025-10-08 |
| 15 | Claude Opus 4.1 | 4.7% | 51.4 | 2025-10-08 |
| 16 | Gemini 2.5 Flash | 4.7% | 52.7 | 2025-10-08 |
| 17 | Claude Sonnet 4 | 4.5% | 50.5 | 2025-10-08 |
| 18 | Claude Sonnet 4thinking | 4.4% | 53.3 | 2025-10-08 |
| 19 |  Nova Premier | 2.0% | 46.1 | 2025-10-08 |
| 20 | Llama 4 Scoutopen ↗ | 1.6% | 38.6 | 2025-10-08 |
| 21 | Llama 4 Maverickopen ↗ | 1.4% | 42.1 | 2025-10-08 |
More multimodal benchmarks · All LLM benchmarks
Related benchmarks
Top head-to-heads on VTB
Cite as: BenchLeader, “VTB leaderboard”, data as of 6 Nov 2026, https://www.benchleader.com/benchmarks/scale_vtb.
VTB: questions
- What does VTB measure?
- Practical vision tasks, from reading charts to spotting differences, graded against references. Scores are reported in percent of tasks solved; higher is better.
- What is the top score on VTB?
- As of 6 Nov 2026, the highest score on VTB is 44.8% by Muse Spark 1.1 (result dated 9 Jul 2026), ahead of GPT-5.4 at 29.2%. Scores come from Scale AI SEAL and are re-read daily.
- Which AI model leads VTB?
- Muse Spark 1.1 leads VTB with 44.8% as of 6 Nov 2026 (result dated 9 Jul 2026), ahead of GPT-5.4 at 29.2%.
- How many models have VTB results?
- 21 model configurations have a VTB result on BenchLeader, all taken from Scale AI SEAL.
- Who runs VTB and how often is it updated?
- VTB is published by Scale AI SEAL. BenchLeader re-reads the published results once a day and records the date each result was published.
- Does VTB count toward the BenchLeader Index?
- No. VTB is shown for reference but left out of the composite index.