Leaderboard
On-device LLM performance rankings powered by Glicko-2
Galaxy Z Fold5
AndroidRank
#115
Rating
1,562
±19 RD
Win Rate
55.9%
Conservative Rating
1,523
TG Rating
1,576
PP Rating
1,522
Matches
740
Record
414W – 326L
Models Tested
| Model | TG Median (tok/s) | PP Median (tok/s) | TG Best | PP Best | Runs |
|---|---|---|---|---|---|
| gemma-3-1b-it.fp16 | 8.42 | 19.37 | 8.42 | 19.37 | 1 |
| qwen2.5-3b-instruct-q5_k_m | 8.07 | 13.48 | 9.86 | 18.08 | 3 |
| gemma-3n-E4B-it-Q4_K_M | 7.28 | 14.42 | 7.28 | 14.42 | 1 |
| gemma-2-2b-it-Q6_K | 6.27 | 23.23 | 6.27 | 23.23 | 1 |
| gemma-3-4b-it-Q8_0 | 5.92 | 14.73 | 5.92 | 14.73 | 1 |
| Llama-3.2-3B-Instruct.Q8_0 | 5.51 | 16.51 | 5.51 | 16.51 | 1 |
| Gemmasutra-Mini-2B-v1-Q6_K | 5.36 | 17.46 | 5.36 | 17.46 | 1 |
| Llama-3.2-3B-Instruct-Q6_K | 5.21 | 9.59 | 6.46 | 12.08 | 3 |
| DeepSeek-R1-Distill-Qwen-1.5B-uncensored.f16 | 4.88 | 8.38 | 4.88 | 8.38 | 1 |
| Phi-4-mini-reasoning-Q8_0 | 3.15 | 8.86 | 3.15 | 8.86 | 1 |
| DeepSeek-R1-Distill-Llama-8B-Q3_K_L | 3.12 | 4.96 | 3.12 | 4.96 | 1 |
| Qwen3-8B-Q6_K | 1.85 | 4.91 | 1.85 | 4.91 | 1 |
Head-to-Head Record
1–50 of 196 rows
1 / 4
Performance by App Version
ImprovedRegressed