Alibaba Cloud · qwen/qwen3-omni-30b-a3b-instruct
| Provider | Alibaba Cloud |
|---|---|
| Family | qwen3-omni |
| Type | vlm |
| Status | active |
| Released | 2025-09-20 |
| Updated | 2025-09-22 |
| Parameters | 35.3B |
| Open weights | yes |
Ordered by predicted speed. Bandwidth sets decode rate; memory decides whether it runs at all.
| Device | Memory | Bandwidth | Best quality | Weights | tok/s | Fastest |
|---|---|---|---|---|---|---|
| Cerebras WSE-3 | 44.0 GB | 21000000.0 GB/s | q6 | 26.44 GB | ~555873.5 | ~833810.3 q4 |
| NVIDIA GB200 Grace Blackwell Superchip | 372.0 GB | 16000.0 GB/s | bf16 | 70.52 GB | ~158.8 | ~635.3 q4 |
| AMD Instinct MI355X | 288.0 GB | 8000.0 GB/s | bf16 | 70.52 GB | ~79.4 | ~317.6 q4 |
| NVIDIA B200 | 180.0 GB | 7700.0 GB/s | bf16 | 70.52 GB | ~76.4 | ~305.7 q4 |
| Google TPU7x (Ironwood) | 192.0 GB | 7380.0 GB/s | bf16 | 70.52 GB | ~73.3 | ~293.0 q4 |
| AMD Instinct MI325X | 256.0 GB | 6000.0 GB/s | bf16 | 70.52 GB | ~59.6 | ~238.2 q4 |
| AMD Instinct MI300X | 192.0 GB | 5300.0 GB/s | bf16 | 70.52 GB | ~52.6 | ~210.4 q4 |
| NVIDIA H200 SXM | 141.0 GB | 4800.0 GB/s | bf16 | 70.52 GB | ~47.6 | ~190.6 q4 |
| NVIDIA H100 NVL | 94.0 GB | 3900.0 GB/s | fp8 | 35.26 GB | ~77.4 | ~154.9 q4 |
| NVIDIA H100 SXM | 80.0 GB | 3350.0 GB/s | fp8 | 35.26 GB | ~66.5 | ~133.0 q4 |
| AMD Instinct MI250X | 128.0 GB | 3200.0 GB/s | bf16 | 70.52 GB | ~31.8 | ~127.1 q4 |
| Google TPU v5p | 95.0 GB | 2765.0 GB/s | bf16 | 70.52 GB | ~27.4 | ~109.8 q4 |
| NVIDIA A100 80GB SXM | 80.0 GB | 2039.0 GB/s | fp8 | 35.26 GB | ~40.5 | ~81.0 q4 |
| NVIDIA H100 PCIe | 80.0 GB | 2000.0 GB/s | fp8 | 35.26 GB | ~39.7 | ~79.4 q4 |
| NVIDIA GeForce RTX 5090 | 32.0 GB | 1792.0 GB/s | q5 | 22.04 GB | ~56.9 | ~71.2 q4 |
| Google TPU v6e (Trillium) | 32.0 GB | 1638.0 GB/s | q5 | 22.04 GB | ~52.0 | ~65.0 q4 |
| AMD Instinct MI210 | 64.0 GB | 1600.0 GB/s | fp8 | 35.26 GB | ~31.8 | ~63.5 q4 |
| NVIDIA A100 40GB SXM | 40.0 GB | 1555.0 GB/s | q6 | 26.44 GB | ~41.2 | ~61.7 q4 |
| Google TPU v4 | 32.0 GB | 1200.0 GB/s | q5 | 22.04 GB | ~38.1 | ~47.6 q4 |
| NVIDIA GeForce RTX 4090 | 24.0 GB | 1008.0 GB/s | q4 | 17.63 GB | ~40.0 | ~40.0 q4 |
| AMD Radeon RX 7900 XTX | 24.0 GB | 960.0 GB/s | q4 | 17.63 GB | ~38.1 | ~38.1 q4 |
| NVIDIA GeForce RTX 3090 | 24.0 GB | 936.0 GB/s | q4 | 17.63 GB | ~37.2 | ~37.2 q4 |
| NVIDIA L40S | 48.0 GB | 864.0 GB/s | fp8 | 35.26 GB | ~17.2 | ~34.3 q4 |
| Apple M3 Ultra | 512.0 GB | 819.0 GB/s | bf16 | 70.52 GB | ~8.1 | ~32.5 q4 |
| Apple M2 Ultra | 192.0 GB | 800.0 GB/s | bf16 | 70.52 GB | ~7.9 | ~31.8 q4 |
| Apple M4 Max | 128.0 GB | 546.0 GB/s | bf16 | 70.52 GB | ~5.4 | ~21.7 q4 |
| Apple M1 Max | 64.0 GB | 400.0 GB/s | fp8 | 35.26 GB | ~7.9 | ~15.9 q4 |
| Apple M2 Max | 96.0 GB | 400.0 GB/s | bf16 | 70.52 GB | ~4.0 | ~15.9 q4 |
| Apple M3 Max | 128.0 GB | 400.0 GB/s | bf16 | 70.52 GB | ~4.0 | ~15.9 q4 |
| NVIDIA L4 | 24.0 GB | 300.0 GB/s | q4 | 17.63 GB | ~11.9 | ~11.9 q4 |
| Apple M4 Pro | 64.0 GB | 273.0 GB/s | fp8 | 35.26 GB | ~5.4 | ~10.8 q4 |
| Apple M4 | 32.0 GB | 120.0 GB/s | q5 | 22.04 GB | ~3.8 | ~4.8 q4 |
Taken from the card's availability section.
| Platform | Id on that platform |
|---|---|
| Hugging Face | Qwen/Qwen3-Omni-30B-A3B-Instruct |
| Benchmark | Evaluated as | Score | Evidence date | Source kind | |
|---|---|---|---|---|---|
| aa_lcr | Qwen3 Omni 30B A3B | 0.0% | 2026-09-10 evaluated | independent evaluator | source |
| critpt | Qwen3 Omni 30B A3B | 0.0% | 2026-09-10 evaluated | independent evaluator | source |
| gpqa_diamond | Qwen3 Omni 30B A3B | 62.02% | 2026-09-10 evaluated | independent evaluator | source |
This card reports no benchmark scores yet.
| Benchmark | Catalogue standing | Score |
|---|