Adobe: firefly-image-4
1 benchmarks
Leaderboard position
#1035
Overall rank (of 1121 models)
-251
Elo rating
1
Benchmarks
Top 92.3%
Percentile
In Adobe (of 2 models)
Above: Kyutai: moshi #1034
Below: DeepSeek: deepseek-llm-67b-chat #1036
Benchmark highlights
Nearest competitor from another company
The model from another company with the most similar benchmark scores.
appeared as nearest in 1 of 1 benchmarks
Benchmarks
Predicted scores
Estimated from 25 similar models and the global Elo rating. These are projections, not measured results.
| Benchmark | Predicted | Confidence | Based on |
|---|---|---|---|
| MATH-500 | ~67.21 | 50% | — |
| AIME 2025 | ~29.83 | 50% | — |
| AIME 2026 | ~52.11 | 50% | — |
| HMMT Feb 2026 | ~43.54 | 50% | — |
| GPQA Diamond | ~33.76 | 50% | — |
| MMLU-Pro | ~34.1 | 50% | — |
| Humanity's Last Exam (no tools) | ~4 | 50% | — |
| IFEval | ~67.17 | 50% | — |
| IFBench | ~35.04 | 50% | — |
| HumanEval | ~41.61 | 50% | — |
| MBPP | ~34.99 | 50% | — |
| Terminal-Bench v2.1 | ~8.6 | 50% | — |
| MMLU | ~39.18 | 50% | — |
| SWE-bench Verified | ~4.51 | 50% | — |
| SWE-bench Pro | ~5.24 | 50% | — |
| SciCode | ~9.66 | 50% | — |
| AA-LCR | ~21.5 | 50% | — |
| Artificial Analysis Intelligence Index v5.0.1 | ~12 | 50% | — |
| Humanity's Last Exam (with tools) | ~11.9 | 50% | — |
| GSM8K | ~41.24 | 50% | — |
| -(CoderPPL) | ~-7.77 | 50% | — |
| LiveCodeBench v6 | ~30.68 | 50% | — |
| MMMU-Pro | ~32.2 | 50% | — |
| MATH | ~14.37 | 50% | — |
| BBH | ~40.8 | 50% | — |