VoxelBench Image-Prompt Leaderboard (VoxelBench Image)
We show this table for reference; we do not rank on it.
A live human-preference benchmark where multimodal models build voxel structures from image references and voters compare anonymous results produced from the same prompt.
Glicko-2 rating on VoxelBench image-prompt leaderboard — November 12, 2026
We mirror the published glicko-2 rating view for VoxelBench image-prompt leaderboard. GPT-6.1 Sol leads the public snapshot at 2260, followed by Gemini 3.1 Pro Preview (1535) and Claude Opus 4.6 (1463). We do not use these results to rank models overall.
GPT-6.1 Sol
OpenAI
1,822 votes · 97.6% win · 95% CI 2123–2397
Gemini 3.1 Pro Preview
2,654 votes · 77.8% win · 95% CI 1421–1649
Claude Opus 4.6
Anthropic
4,971 votes · 63.2% win · 95% CI 1379–1547
10 modelsMultimodal & GroundedCurrentDisplay onlyUpdated November 12, 2026
Glicko-2 rating table (10 models)
| Rank | Model / configuration | Score | Parameters (B) | Open / closed |
|---|---|---|---|---|
| 1 | GPT-6.1 SolOpenAI1,822 votes · 97.6% win · 95% CI 2123–2397 | 2260 | Not reported | Closed |
| 2 | Gemini 3.1 Pro PreviewGoogle2,654 votes · 77.8% win · 95% CI 1421–1649 | 1535 | Not reported | Unknown |
| 3 | Claude Opus 4.6Anthropic4,971 votes · 63.2% win · 95% CI 1379–1547 | 1463 | Not reported | Closed |
| 4 | GPT-5.4OpenAI3,570 votes · 62.0% win · 95% CI 1341–1537 | 1439 | Not reported | Closed |
| 5 | Gemini 3 Pro PreviewGoogle5,991 votes · 64.3% win · 95% CI 1324–1492 | 1408 | Not reported | Unknown |
| 6 | GPT-5.2 (xHigh)OpenAI5,909 votes · 61.1% win · 95% CI 1269–1437 | 1353 | Not reported | Unknown |
| 7 | Claude Opus 4.5Anthropic5,915 votes · 53.2% win · 95% CI 1238–1402 | 1320 | Not reported | Closed |
| 8 | Claude Opus 4.7Anthropic5,322 votes · 46.9% win · 95% CI 1192–1360 | 1276 | Not reported | Closed |
| 9 | Gemini 2.5 ProGoogle4,468 votes · 41.8% win · 95% CI 1096–1288 | 1192 | Not reported | Closed |
| 10 | GPT-5 (high)OpenAI4,672 votes · 43.7% win · 95% CI 1074–1258 | 1166 | Not reported | Closed |
How to read this leaderboard
The board measures which rendered voxel builds voters prefer when models receive an image reference. Higher ratings indicate stronger pairwise results against the tested field, not absolute reconstruction accuracy.
Operator receipt: 10 sourced rows are currently displayable on this page; the leading published row is GPT-6.1 Sol at 2260.
Honest limit: The source does not publish a fixed, versioned task set or a controlled reconstruction metric. Ratings can move as prompts, models, and voters change, and the complete generation pipeline contributes to every result.
How we show VoxelBench image-prompt ratings
We mirror the image-prompt table from VoxelBench's public leaderboard API. The source page shows models after at least 50 votes and ranks the eligible rows with Glicko-2; the snapshot keeps each rating, deviation, 95% confidence interval, vote count, win rate, and win/loss/tie record.
Voters compare two anonymous voxel structures produced for the same image-reference prompt. The result mixes model behavior with VoxelBench's generation tools, instructions, settings, prompt mix, renderer, and voter pool, so we keep it display only and separate from weighted model rankings.
Snapshot
The published VoxelBench Image snapshot places GPT-6.1 Sol first at 2260. The third row is 797 score units behind. The broader top-10 range is 1094 score units, so the table still separates the published systems.
10 models have been evaluated on VoxelBench Image. The benchmark falls in the Multimodal & Grounded category. We keep external benchmark mirrors separate from the weighted global scoring system, so these results remain source-specific evidence. VoxelBench Image is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.
About VoxelBench Image
Year
2025
Tasks
Live image-reference prompts for 3D voxel construction
Format
Glicko-2 rating from blind pairwise votes
Difficulty
Visual grounding, 3D construction, and aesthetic quality
The mirrored image-prompt view keeps only models with at least 50 votes, matching the public page. Each row preserves the Glicko-2 rating, deviation, confidence interval, vote count, win rate, and win/loss/tie record.
Freshness and provenance
Version
Live VoxelBench Glicko-2
Refresh cadence
Rolling
Staleness state
Current
Question availability
Prompts browsable; full task set not versioned
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 VoxelBench Image measure?
A live human-preference benchmark where multimodal models build voxel structures from image references and voters compare anonymous results produced from the same prompt.
Which model leads the published VoxelBench Image snapshot?
GPT-6.1 Sol currently leads the published VoxelBench Image snapshot with 2260 glicko-2 rating. BenchLM shows this benchmark for display only and does not use it in overall rankings.
How many models are evaluated on VoxelBench Image?
The November 12, 2026 snapshot contains 10 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.