Measured over the last 30 days
- Models Tracked
- 55
- Avg Tokens / Second
- 17.70
- Avg Time to First Token (ms)
- 9279.82
- Last Updated
- Nov 8, 2026
All deepinfra models
55 live · one row per model · fastest first · 2 retired lanes not listed · 10 folded behind their current lane| Model | Avg Toks/Sec | Min | Max | Avg TTF (ms) |
|---|---|---|---|---|
| Nemotron 3.5 Lightning | 77.30 | 77.30 | 77.30 | 540.00 |
| Llama 3 8B Lunaris | 58.30 | 58.30 | 58.30 | 750.00 |
| Gemma 4 26B A4B | 44.20 | 29.90 | 58.40 | 700.00 |
| Qwen3 Coder 480B A35B | 38.30 | 38.30 | 38.30 | 650.00 |
| Llama 3.1 Euryale 70B v2.2 | 35.60 | 35.60 | 35.60 | 540.00 |
| Qwen3 Next 80B A3B Instruct | 34.20 | 9.33 | 59.10 | 3440.00 |
| MythoMax 13B | 32.40 | 30.90 | 34.00 | 510.00 |
| Mistral Small 3.2 24B | 31.50 | 30.30 | 32.60 | 690.00 |
| Gemma 3 27B | 31.30 | 30.30 | 32.20 | 620.00 |
| Phi 4 | 31.30 | 8.52 | 54.00 | 3750.00 |
| GLM 5.2 | 29.10 | 14.30 | 43.30 | 1890.00 |
| Llama 4 Scout | 27.90 | 26.20 | 29.70 | 720.00 |
| Mistral Small 3 | 26.30 | 24.00 | 28.60 | 1000.00 |
| Llama 3.1 70B Instruct | 25.20 | 25.20 | 25.20 | 1010.00 |
| Llama 3.1 8B Instruct | 23.90 | 23.90 | 23.90 | 680.00 |
| Gemma 3 4B | 23.50 | 23.50 | 23.50 | 1030.00 |
| Qwen3 VL 30B A3B Instruct | 22.60 | 19.50 | 25.70 | 630.00 |
| Mistral Nemo | 22.40 | 20.00 | 24.80 | 540.00 |
| Gemma 4 31B | 21.80 | 7.10 | 46.30 | 1090.00 |
| Granite 4.2 8B | 21.70 | 21.70 | 21.70 | 2450.00 |
| DeepSeek V3.2 | 19.20 | 8.03 | 33.20 | 1600.00 |
| Step 5 Preview | 19.10 | 19.10 | 19.10 | 3120.00 |
| DeepSeek V4 Flash 0731 | 18.50 | 16.10 | 20.80 | 2570.00 |
| Inkling Small | 18.50 | 6.22 | 30.80 | 5680.00 |
| Nemotron 3.5 Super VL | 18.10 | 0.88 | 38.40 | 8580.00 |
| Nemotron 3 Ultra | 16.60 | 16.60 | 16.60 | 3020.00 |
| gpt-oss-120b | 15.40 | 2.61 | 29.80 | 6690.00 |
| Gemma 3 12B | 14.90 | 12.90 | 16.80 | 2360.00 |
| MiMo-V2.6-Flash | 14.90 | 1.30 | 25.00 | 9690.00 |
| Llama 3.3 70B Instruct | 14.40 | 1.67 | 25.20 | 3440.00 |
| Qwen3 235B A22B Instruct 2507 | 13.20 | 5.61 | 22.40 | 1720.00 |
| Qwen3 14B | 12.90 | 12.90 | 12.90 | 4540.00 |
| MiniMax M3 | 11.50 | 3.12 | 24.50 | 6680.00 |
| Qwen3 30B A3B | 11.10 | 10.30 | 11.90 | 5370.00 |
| DeepSeek V3.1 | 9.89 | 6.82 | 13.00 | 810.00 |
| Qwen3 VL 235B A22B Instruct | 9.72 | 8.34 | 11.10 | 1140.00 |
| Qwen3.5-35B-A3B | 8.92 | 8.92 | 8.92 | 6880.00 |
| DeepSeek V3 | 7.95 | 7.95 | 7.95 | 4490.00 |
| DeepSeek V4.1 Flash 1017 | 6.97 | 4.52 | 10.30 | 10500.00 |
| Muse Glimmer 30B | 6.76 | 6.76 | 6.76 | 8970.00 |
| Nemotron 3 Super | 6.07 | 5.02 | 7.12 | 10200.00 |
| Qwen3.6 35B A3B | 4.54 | 1.08 | 7.38 | 20600.00 |
| Ling 3.1 Flash | 4.45 | 3.49 | 5.41 | 12900.00 |
| Ling 3.0 Flash | 4.29 | 4.29 | 4.29 | 15200.00 |
| R1 0528 | 3.37 | 1.92 | 4.32 | 18900.00 |
| Qwen3.5 397B A17B | 3.37 | 1.49 | 6.38 | 21100.00 |
| Llama Guard 4 12B | 3.17 | 3.17 | 3.17 | 910.00 |
| Qwen3 32B | 3.17 | 3.17 | 3.17 | 17600.00 |
| Kimi K2.6 | 2.94 | 2.12 | 4.54 | 20900.00 |
| DeepSeek V4.1 Flash | 2.84 | 2.73 | 2.95 | 21900.00 |
| Qwen3.5-27B | 2.31 | 2.31 | 2.31 | 26300.00 |
| GLM 4.6 | 1.61 | 0.83 | 2.40 | 39100.00 |
| Qwen2.5 72B Instruct | 1.58 | 1.58 | 1.58 | 38300.00 |
| GLM 4.7 | 1.49 | 1.29 | 1.82 | 41200.00 |
| Qwen3.5-9B | 0.74 | 0.74 | 0.74 | 84200.00 |
Frequently Asked Questions
Which deepinfra model is fastest?
Based on recent tests, Nemotron 3.5 Lightning shows the highest average throughput among tracked deepinfra models.
How many recent measurements feed this dashboard?
This provider summary aggregates 3846 individual prompts measured across 3243 monitoring runs over the past month.