Verified on RTX 3090
This page tracks model tags with measured local runs on RTX 3090. Use it when you want evidence-backed local performance instead of pure baseline estimates.
Measured model tags
| Model | Tag | Tokens/s | Latency (ms) | Test time | Detail |
|---|---|---|---|---|---|
| GPT-OSS 20B CLOUD | gpt-oss:20b | 156.1 | 1524 | 2026-04-29T05:39:58Z | Open |
| Qwen3 Coder 30B CLOUD | qwen3-coder:30b | 140.5 | 935 | 2026-06-17T07:31:11Z | Open |
| Qwen3 8B FP16 | qwen3:8b | 121.7 | 1429 | 2026-06-17T07:31:11Z | Open |
| Qwen2.5 Coder 32B FP16 | qwen2.5-coder:32b | 92.2 | 1609 | 2026-06-17T07:31:11Z | Open |
| Qwen2.5 14B FP16 | qwen2.5:14b | 84.0 | 946 | 2026-04-29T05:39:58Z | Open |
| Ministral 3 14B FP16 | ministral-3:14b | 79.2 | 2047 | 2026-06-17T07:31:11Z | Open |
| DeepSeek-R1 14B FP16 | deepseek-r1:14b | 74.4 | 2119 | 2026-06-17T07:31:11Z | Open |
| Nemotron 3 Nano 30B FP16 | nemotron-3-nano:30b | 57.0 | 2468 | 2026-04-01T11:53:50Z | Open |
| Mistral Small 22B FP16 | mistral-small:22b | 54.8 | 1937 | 2026-06-17T07:31:11Z | Open |
| Qwen3.6 35B FP16 | qwen3.6:35b | 47.8 | 2628 | 2026-06-17T07:31:11Z | Open |
| Translategemma 27B FP16 | translategemma:27b | 41.3 | 3142 | 2026-04-01T11:53:50Z | Open |
| Gemma 3 27B FP16 | gemma3:27b | 39.5 | 3136 | 2026-06-17T07:31:11Z | Open |
| QwQ 32B FP16 | qwq:32b | 35.8 | 3032 | 2026-06-17T07:31:11Z | Open |
| Qwen3.5 35B FP16 | qwen3.5:35b | 33.8 | 3424 | 2026-06-17T07:31:11Z | Open |
| Glm 4.7 Flash 7B FP16 | glm-4.7-flash:bf16 | 11.2 | 9291 | 2026-03-04T09:01:38Z | Open |
| Llama 4 16X17B FP16 | llama4:16x17b | 9.1 | 7779 | 2026-06-10T06:45:58Z | Open |
| Qwen3.5 122B FP16 | qwen3.5:122b | 4.9 | 11915 | 2026-02-26T19:19:16Z | Open |
| Llama 3.3 70B FP16 | llama3.3:70b | 3.5 | 15676 | 2026-04-29T05:39:58Z | Open |
Validation notes
- Measured rows come from actual benchmark snapshots, not template placeholders.
- Catalog baseline values stay visible on model pages for comparison.
- For heavy profiles, use cloud fallback links after local saturation.