Measured over the last 30 days
- Models Tracked
- 16
- Avg Tokens / Second
- 27.99
- Avg Time to First Token (ms)
- 4303.13
- Last Updated
- Oct 14, 2026
All coreweave models
16 live · one row per model · fastest first · 6 folded behind their current lane| Model | Avg Toks/Sec | Min | Max | Avg TTF (ms) |
|---|---|---|---|---|
| Llama 3.1 8B Instruct | 80.70 | 80.70 | 80.70 | 400.00 |
| GLM 5.2 | 53.10 | 8.71 | 70.00 | 1620.00 |
| Gemma 4 26B A4B | 50.90 | 20.80 | 72.70 | 510.00 |
| Llama 3.3 70B Instruct | 45.60 | 10.50 | 58.50 | 700.00 |
| DeepSeek V3.1 | 42.80 | 41.10 | 44.40 | 680.00 |
| Kimi K2.7 Code | 25.80 | 8.80 | 46.80 | 2800.00 |
| MiniMax M3 | 23.80 | 3.93 | 43.40 | 3560.00 |
| gpt-oss-20b | 23.40 | 0.92 | 89.20 | 3690.00 |
| Granite 4.2 8B | 21.80 | 16.70 | 25.80 | 2440.00 |
| Gemma 4 31B | 21.40 | 3.56 | 42.30 | 2030.00 |
| DeepSeek V4 Flash 0731 | 15.10 | 1.09 | 37.50 | 6880.00 |
| Nemotron 3.5 Lightning | 12.30 | 12.30 | 12.30 | 4880.00 |
| gpt-oss-120b | 10.80 | 5.48 | 14.10 | 5410.00 |
| Kimi K2.6 | 9.51 | 2.33 | 34.10 | 9850.00 |
| Qwen3.6 35B A3B | 5.65 | 3.90 | 7.17 | 11200.00 |
| DeepSeek V4.1 Flash | 5.16 | 4.72 | 5.61 | 12200.00 |
Frequently Asked Questions
Which coreweave model is fastest?
Based on recent tests, Llama 3.1 8B Instruct shows the highest average throughput among tracked coreweave models.
How many recent measurements feed this dashboard?
This provider summary aggregates 2867 individual prompts measured across 2441 monitoring runs over the past month.