Measured over the last 30 days
- Models Tracked
- 21
- Avg Tokens / Second
- 14.28
- Avg Time to First Token (ms)
- 13453.93
- Last Updated
- Sep 1, 2026
All phala models
21 tracked · fastest first| Model | Avg Toks/Sec | Min | Max | Avg TTF (ms) |
|---|---|---|---|---|
| Qwen2.5 7B Instruct | 52.40 | 39.70 | 59.70 | 610.00 |
| Qwen2.5 7B Instruct | 46.30 | 28.20 | 62.00 | 730.00 |
| Gemma 4 31B | 28.00 | 3.44 | 51.40 | 1210.00 |
| gpt-oss-120b | 21.90 | 14.80 | 30.80 | 2590.00 |
| GLM 5.2 | 15.40 | 8.68 | 20.10 | 3310.00 |
| Gemma 3 27B | 12.90 | 6.80 | 18.90 | 1090.00 |
| DeepSeek V4 Flash 0731 | 11.80 | 7.87 | 14.10 | 4810.00 |
| gpt-oss-20b | 10.70 | 10.70 | 10.70 | 4910.00 |
| Muse Glimmer 30B | 6.31 | 5.95 | 6.73 | 9120.00 |
| Muse Glimmer 30B | 6.28 | 5.33 | 8.13 | 9350.00 |
| Qwen3.6 35B A3B | 5.57 | 4.05 | 6.91 | 11600.00 |
| DeepSeek V3.2 | 5.25 | 2.96 | 7.27 | 2660.00 |
| Hy3 | 4.67 | 4.67 | 4.67 | 13300.00 |
| Kimi K2.5 | 4.56 | 1.50 | 7.34 | 20300.00 |
| Kimi K2.6 | 3.62 | 2.22 | 5.58 | 18200.00 |
| Kimi K3 | 3.03 | 1.68 | 4.01 | 23200.00 |
| DeepSeek V4 Flash 0423 | 2.66 | 2.41 | 2.91 | 4670.00 |
| Qwen3.5 397B A17B | 1.90 | 0.95 | 2.96 | 34200.00 |
| GLM 5.1 | 1.45 | 1.15 | 1.96 | 45300.00 |
| Qwen3.5-27B | 0.96 | 0.67 | 1.50 | 68000.00 |
| DeepSeek V4 Pro 0813 | 0.20 | 0.20 | 0.20 | 45500.00 |
Frequently Asked Questions
Which phala model is fastest?
Based on recent tests, Qwen2.5 7B Instruct shows the highest average throughput among tracked phala models.
How many recent measurements feed this dashboard?
This provider summary aggregates 5495 individual prompts measured across 4547 monitoring runs over the past month.