Scale Labs Visual Tool Bench (Visual Tool Bench)
We show this table for reference; we do not rank on it.
A Scale Labs public leaderboard mirrored as display-only reference data. It does not affect BenchLM rankings.
Scale score on Visual Tool Bench — November 12, 2026 snapshot
We mirror the published scale score view for Visual Tool Bench. Muse Spark 1.1 leads the public snapshot at 44.8%, followed by gpt-5.4-2026-03-05 (reasoning effort = high) (29.2%) and gemini-3.1-pro-preview (29.0%). We do not use these results to rank models overall.
Muse Spark 1.1
Meta
gpt-5.4-2026-03-05 (reasoning effort = high)
OpenAI
gemini-3.1-pro-preview
21 modelsAgenticCurrentDisplay onlyUpdated November 12, 2026 snapshot
Scale score table (21 models)
| Rank | Model / configuration | Score | Parameters (B) | Open / closed |
|---|---|---|---|---|
| 1 | Muse Spark 1.1Meta | 44.8% | Not reported | Closed |
| 2 | gpt-5.4-2026-03-05 (reasoning effort = high)OpenAI | 29.2% | Not reported | Unknown |
| 3 | gemini-3.1-pro-previewGoogle | 29.0% | Not reported | Unknown |
| 4 | Claude Opus 4.6 (Adaptive)Anthropic | 27.5% | Not reported | Closed |
| 5 | gemini-3-pro-previewGoogle | 26.9% | Not reported | Unknown |
| 6 | gpt-5-2025-08-07-thinkingOpenAI | 18.7% | Not reported | Unknown |
| 7 | gpt-5-2025-08-07OpenAI | 17.0% | Not reported | Unknown |
| 8 | o3OpenAI | 13.7% | Not reported | Closed |
| 9 | gemini-2.5-pro-preview-06-05Google | 11.8% | Not reported | Unknown |
| 10 | o4-miniOpenAI | 11.1% | Not reported | Closed |
| 11 | Claude Sonnet 4.5 ThinkingAnthropic | 6.2% | Not reported | Closed |
| 12 | Claude Sonnet 4.5Anthropic | 5.6% | Not reported | Closed |
| 13 | GPT-4.1OpenAI | 5.5% | Not reported | Closed |
| 14 | claude-opus-4-1-20250805-thinkingAnthropic | 5.2% | Not reported | Unknown |
| 15 | claude-opus-4-1-20250805Anthropic | 4.7% | Not reported | Unknown |
| 16 | Gemini 2.5 FlashGoogle | 4.7% | Not reported | Closed |
| 17 | claude-sonnet-4Anthropic | 4.5% | Not reported | Unknown |
| 18 | claude-sonnet-4-thinkingAnthropic | 4.4% | Not reported | Unknown |
| 19 | nova-premierAmazon | 2% | Not reported | Unknown |
| 20 | llama4-scoutMeta | 1.6% | Not reported | Unknown |
| 21 | llama4-maverickMeta | 1.4% | Not reported | Unknown |
How to read this leaderboard
Compare the published configurations as complete evaluation systems. The source can combine a base model, agent scaffold, tools, budget, and inference setting in each result.
Operator receipt: 21 sourced rows are currently displayable on this page; the leading published row is Muse Spark 1.1 at 44.8%.
Honest limit: This Scale table is display-only context, not benchmark provenance or a weighted model-only comparison.
How BenchLM shows Visual Tool Bench
BenchLM mirrors 21 published rows from Scale Labs’ public Visual Tool Bench leaderboard, captured on November 12, 2026 snapshot.
The table is display only. It is useful context for a published agent or model configuration, but it does not enter BenchLM’s overall or category rankings.
The published Visual Tool Bench snapshot places Muse Spark 1.1 first at 44.8%. The third row is 15.8 points behind. The broader top-10 range is 33.7 points, so the table still separates the published systems.
21 models have been evaluated on Visual Tool Bench. The benchmark falls in the Agentic category. We keep external benchmark mirrors separate from the weighted global scoring system, so these results remain source-specific evidence. Visual Tool Bench is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.
About Visual Tool Bench
Year
2026
Tasks
21 published rows
Format
Published Scale leaderboard score
Difficulty
External agent and model evaluation
BenchLM mirrors 21 published rows from the Visual Tool Bench public table captured on November 12, 2026 snapshot.
Freshness and provenance
Version
Visual Tool Bench 2026
Refresh cadence
Quarterly
Staleness state
Current
Question availability
Public benchmark set
BenchLM uses freshness metadata to decide whether a benchmark should still be treated as a strong differentiator, a benchmark to watch, or a display-only reference. For the full scoring policy, see the BenchLM methodology page.
Questions
What does Visual Tool Bench measure?
A Scale Labs public leaderboard mirrored as display-only reference data. It does not affect BenchLM rankings.
Which model leads the published Visual Tool Bench snapshot?
Muse Spark 1.1 currently leads the published Visual Tool Bench snapshot with 44.8% scale score. BenchLM shows this benchmark for display only and does not use it in overall rankings.
How many models are evaluated on Visual Tool Bench?
The November 12, 2026 snapshot snapshot contains 21 AI models.
Know when it’s worth switching models
The model to choose, the cheaper alternative, and the release we would wait on.
Read a sample issueJoin 7,000+ subscribers.
One email each week. Unsubscribe anytime.