Qwen
Benchmarks
TaskDatasetSOTA resultTrendResults
Qwen3-1.7B Small Target
98Similarity Score
18
Qwen Multi-task Evaluation Suite 2.5 (test)
59.5MC Score
18
Qwen 3 1.7B
0.149Time Reduction
18
Qwen 7B 2.5
0.998Fingerprint Similarity Score
18
Qwen-4B Target (Transferability set)
50ASR
17
Qwen (test)
100Detection Accuracy
15
Qwen
83.31Recall@k
15
Qwen2.5-3B-Instruct K3V3 deployment
155.7KV Cache Size (MB)
13
Qwen2.5-3B-Instruct K3V3
17.84Median Decode Latency (ms/step)
13
Qwen3-0.6B
7.03Compression Ratio
13
Qwen 0.5B -> 3B 2.5
0.32Arithmetic (-) Score
13
Qwen-32B
80.81MMLU
13
Qwen1.5-MoE (test)
100Watermark Success Rate (WSR)
12
Qwen 72B 2.5
0.016Latency Overhead
12
Qwen3 Context length (50K)
6.02Throughput Speedup (α)
12
Qwen3 Context length 10K
2.76Throughput Speedup (α)
12
Qwen1.5-MoE-A2.7B evaluation suite zero-shot
6.79PPL
11
Qwen3-4B Calibration
0Brier Delta
10
Qwen 14B 2.5
100AUC
10
Qwen 7B 2.5
100AUC
10
Qwen2-VL
96.4ASR
10
Qwen3 1.7B
100Mean Certifiable Acceptance Length (epsilon=0.01)
9
Qwen3-4B base (train)
129.8Per-step Latency (ms)
9
Qwen 1.5B Instruct 2.5
67.47TPS
9
Qwen-3-4B
11.08Input Success Rate
9