Measured over the last 30 days
- Models Tracked
- 42
- Avg Tokens / Second
- 16.51
- Avg Time to First Token (ms)
- 10091.90
- Last Updated
- Oct 17, 2026
All alibaba models
42 live · one row per model · fastest first · 13 retired lanes not listed · 6 folded behind their current lane| Model | Avg Toks/Sec | Min | Max | Avg TTF (ms) |
|---|---|---|---|---|
| Qwen3 VL 8B Instruct | 67.80 | 67.80 | 67.80 | 400.00 |
| Qwen3.5-Flash | 56.90 | 56.90 | 56.90 | 810.00 |
| Qwen3 VL 30B A3B Instruct | 56.10 | 56.10 | 56.10 | 670.00 |
| Qwen3 30B A3B Instruct 2507 | 46.20 | 46.20 | 46.20 | 630.00 |
| Qwen3 Coder Flash | 45.10 | 45.10 | 45.10 | 900.00 |
| Qwen3 Next 80B A3B Instruct | 41.30 | 41.30 | 41.30 | 760.00 |
| Qwen3 Coder Next | 35.90 | 35.90 | 35.90 | 1350.00 |
| Qwen-Plus | 33.40 | 33.40 | 33.40 | 780.00 |
| Qwen3 235B A22B Instruct 2507 | 32.70 | 30.40 | 35.20 | 910.00 |
| Qwen3 Coder 480B A35B | 28.20 | 28.10 | 28.20 | 1210.00 |
| DeepSeek V3.2 | 24.20 | 21.40 | 27.00 | 1170.00 |
| Qwen3 VL 235B A22B Instruct | 24.20 | 24.20 | 24.20 | 1030.00 |
| Qwen3 Coder 30B A3B Instruct | 18.70 | 18.70 | 18.70 | 2910.00 |
| DeepSeek V4 Flash 0731 | 15.70 | 3.31 | 27.20 | 4540.00 |
| Qwen3 30B A3B | 14.70 | 14.70 | 14.70 | 3660.00 |
| Qwen3 VL 30B A3B Thinking | 13.30 | 13.30 | 13.30 | 4360.00 |
| Qwen3.8 Max Prime | 11.10 | 4.12 | 17.30 | 5100.00 |
| Kimi K2.7 Code | 10.60 | 4.28 | 15.50 | 6990.00 |
| DeepSeek V4 Flash 0423 | 10.50 | 8.86 | 12.90 | 5630.00 |
| DeepSeek V4 Pro 0423 | 10.50 | 3.61 | 14.10 | 5590.00 |
| Qwen3 Next 80B A3B Thinking | 8.65 | 8.65 | 8.65 | 7000.00 |
| Qwen3 14B | 8.63 | 8.63 | 8.63 | 8070.00 |
| GLM 5.2 | 7.76 | 2.95 | 35.40 | 11900.00 |
| Qwen3 235B A22B Thinking 2507 | 7.34 | 7.34 | 7.34 | 7480.00 |
| Qwen3.5-35B-A3B | 7.30 | 7.30 | 7.30 | 8310.00 |
| Qwen3.8 Omni Flash | 7.11 | 1.19 | 16.30 | 13200.00 |
| Qwen3.5-122B-A10B | 6.97 | 6.82 | 7.13 | 9030.00 |
| Qwen3.5-27B | 5.24 | 5.24 | 5.24 | 11300.00 |
| qwen3.8-max-0902 | 5.20 | 3.04 | 8.87 | 12100.00 |
| Qwen3.6 Flash | 4.33 | 4.33 | 4.33 | 14300.00 |
| Qwen3.7 Flash | 3.75 | 3.75 | 3.75 | 16600.00 |
| Qwen3.5 397B A17B | 3.61 | 2.39 | 5.31 | 17600.00 |
| Qwen3.8 2.4T A95B | 3.19 | 3.19 | 3.19 | 19200.00 |
| DeepSeek V4.1 Flash | 2.97 | 2.97 | 2.97 | 21500.00 |
| Qwen3.5 Plus 2026-02-15 | 2.69 | 2.69 | 2.69 | 23400.00 |
| Qwen3.6 Plus | 2.61 | 2.61 | 2.61 | 24400.00 |
| GLM 5.1 | 2.50 | 1.62 | 3.49 | 25700.00 |
| Qwen3.5 Plus 2026-04-20 | 2.18 | 2.18 | 2.18 | 29200.00 |
| Qwen3.7 Plus | 1.98 | 1.98 | 1.98 | 30700.00 |
| DeepSeek V4 Pro 0813 | 1.76 | 1.76 | 1.76 | 35700.00 |
| GLM 5.3 | 0.25 | 0.05 | 4.96 | 19500.00 |
| GLM 5.3 Prime | 0.25 | 0.17 | 0.37 | 8270.00 |
Frequently Asked Questions
Which alibaba model is fastest?
Based on recent tests, Qwen3 VL 8B Instruct shows the highest average throughput among tracked alibaba models.
How many recent measurements feed this dashboard?
This provider summary aggregates 2112 individual prompts measured across 1966 monitoring runs over the past month.