Apple: simplesd-4b-instruct
2 benchmarks
Leaderboard position
#822
Overall rank (of 1121 models)
419
Elo rating
2
Benchmarks
Top 73.3%
Percentile
In Apple (of 13 models)
Above: Anthropic: claude-3-haiku #821
Below: Apple: openelm-3b-instruct #823
Benchmark highlights
Best benchmark
LiveCodeBench v5
45.7
Worst benchmark
LiveCodeBench v6
41.5
Nearest competitor from another company
The model from another company with the most similar benchmark scores.
appeared as nearest in 1 of 2 benchmarks
1/2
50.0%
1/2
50.0%
| Model | Votes | % of benchmarks |
|---|---|---|
| Xiaomi: MiMo-V2.5 Pro | 1 | 50.0% |
| Mistral AI: Magistral Small 1.2 | 1 | 50.0% |
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 | ~78.83 | 50% | — |
| AIME 2025 | ~51.82 | 50% | — |
| AIME 2026 | ~65.07 | 50% | — |
| HMMT Feb 2026 | ~56.55 | 50% | — |
| GPQA Diamond | ~48.5 | 50% | — |
| MMLU-Pro | ~50.54 | 50% | — |
| Humanity's Last Exam (no tools) | ~4 | 50% | — |
| IFEval | ~76.33 | 50% | — |
| IFBench | ~46.13 | 50% | — |
| HumanEval | ~61.07 | 50% | — |
| MBPP | ~60.48 | 50% | — |
| Terminal-Bench v2.1 | ~8.6 | 50% | — |
| MMLU | ~56.45 | 50% | — |
| SWE-bench Verified | ~25.53 | 50% | — |
| SWE-bench Pro | ~5.4 | 50% | — |
| SciCode | ~20.12 | 50% | — |
| AA-LCR | ~34.59 | 50% | — |
| Artificial Analysis Intelligence Index v5.0.1 | ~12 | 50% | — |
| Humanity's Last Exam (with tools) | ~11.9 | 50% | — |
| GSM8K | ~66.44 | 50% | — |
| -(CoderPPL) | ~-3.62 | 50% | — |
| MMMU-Pro | ~44.78 | 50% | — |
| MATH | ~42.49 | 50% | — |
| BBH | ~55.67 | 50% | — |
| ARC-Easy | ~71.39 | 50% | — |