Model
GPT-4V
OpenAI · closed weights
Math
MathAgentic / Tool Use
Agentic / Tool UseMultimodal
Multimodal| Benchmark | Score | Metric | Harness | Source | Confidence | Link |
|---|---|---|---|---|---|---|
| BenchLMM | 58.4 | gpt-3.5_score | N/A | Independent | medium | arxiv.org |
| CORE-MM | 77.9 | abductive | N/A | Independent | medium | arxiv.org |
| EmbSpatial-Bench | 36.1 | generation | N/A | Independent | medium | arxiv.org |
| HallusionBench | 12.2 | question_pair_acc | N/A | Independent | medium | arxiv.org |
| InfiMM-Eval | 74.4 | overall_score | N/A | Independent | medium | arxiv.org |
| MM-Vet | 67.7 | gpt-4_score | N/A | Independent | medium | arxiv.org |
| REBUS | 24.0 | accuracy | N/A | Independent | medium | arxiv.org |
| Winoground | 54.5 | text_score | N/A | Independent | medium | arxiv.org |
Long Context
Long ContextSafety & Alignment
Safety & Alignment| Benchmark | Score | Metric | Harness | Source | Confidence | Link |
|---|---|---|---|---|---|---|
| AutoHallusion | 66.0 | overall_accuracy | N/A | Independent | medium | arxiv.org |
Reproduced by the community
Loading community runs…