Measured over the last 30 days
- Models Tracked
- 19
- Avg Tokens / Second
- 12.46
- Avg Time to First Token (ms)
- 9553.68
- Last Updated
- Oct 14, 2026
All atlascloud models
19 live · one row per model · fastest first · 1 folded behind their current lane| Model | Avg Toks/Sec | Min | Max | Avg TTF (ms) |
|---|---|---|---|---|
| DeepSeek V3.2 Exp | 35.00 | 35.00 | 35.00 | 1390.00 |
| Qwen3 235B A22B Instruct 2507 | 23.30 | 23.30 | 23.30 | 1690.00 |
| DeepSeek V3.1 Terminus | 23.10 | 23.10 | 23.10 | 1870.00 |
| DeepSeek V3.2 | 22.60 | 17.90 | 28.30 | 1450.00 |
| GLM 5.2 | 19.60 | 13.70 | 25.40 | 2440.00 |
| GLM 5.1 | 18.90 | 15.90 | 23.40 | 2280.00 |
| DeepSeek V4 Flash 0731 | 16.90 | 10.80 | 22.90 | 4040.00 |
| MiniMax M3 | 15.80 | 7.38 | 22.60 | 4070.00 |
| MiMo-V2.5-Pro | 11.70 | 11.70 | 11.70 | 3810.00 |
| DeepSeek V4 Flash 0423 | 9.08 | 2.13 | 11.10 | 7570.00 |
| MiniMax M2.5 | 8.82 | 4.00 | 19.50 | 9070.00 |
| Kimi K2.5 | 6.83 | 4.87 | 8.80 | 8940.00 |
| Qwen3.5-35B-A3B | 4.42 | 4.42 | 4.42 | 14300.00 |
| LongCat 2.0 | 4.36 | 3.25 | 5.47 | 14800.00 |
| Kimi K2.6 | 3.90 | 2.69 | 5.91 | 17000.00 |
| Qwen3.5 397B A17B | 3.54 | 2.97 | 4.24 | 17700.00 |
| Qwen3.5-122B-A10B | 3.53 | 3.53 | 3.53 | 18000.00 |
| Qwen3.5-27B | 2.97 | 2.97 | 2.97 | 20400.00 |
| MiniMax M2.7 | 2.43 | 1.50 | 3.16 | 30700.00 |
Frequently Asked Questions
Which atlascloud model is fastest?
Based on recent tests, DeepSeek V3.2 Exp shows the highest average throughput among tracked atlascloud models.
How many recent measurements feed this dashboard?
This provider summary aggregates 2174 individual prompts measured across 1959 monitoring runs over the past month.