Manufacturer

Qwen (Alibaba)

Qwen-Modellfamilie von Alibaba.

Website ↗ 🤗 Hugging Face Filter in leaderboard

33Models

1549Benchmarks

1537Performance

12Harness

Best token generation

1.716,6 tok/s

Qwen3-Coder-30B-A3B-Instruct

Best prefill

112.346,4 tok/s

Qwen3-0.6B

∅ Token generation

238,9 tok/s

Average of 1537 runs

∅ Prefill

5.934,9 tok/s

Average of 1537 runs

Best TTFT

121 ms

Qwen3-0.6B

Best harness rate

100%

Qwen3.8-27B-Uncensored

🏆

Qwen3-Coder-30B-A3B-Instruct · fastest model at 1.716,6 tok/s token generation
Best prefill of this manufacturer: 112.346,4 tok/s (Qwen3-0.6B)

Top models by token generation

Best measured generation throughput per model (tok/s), published performance benchmarks.

Qwen3-Coder-30B-A3B-Instruct

1.716,6 tok/s

Qwen3-30B-A3B-Thinking-2507

1.683,8 tok/s

Qwen3-Omni-30B-A3B-Thinking

1.673,6 tok/s

Qwen3-30B-A3B-Instruct-2507

1.671,0 tok/s

Qwen3-VL-30B-A3B-Instruct

1.655,9 tok/s

Qwen3.8-Flash-Next

1.564,6 tok/s

Qwen3.5-35B-A3B

1.520,4 tok/s

Qwen3.6-35B-A3B

1.489,1 tok/s

Models

33 assigned

All benchmarks →

Huihui-Qwen3-VL-30B-A3B-Instruct-abliterated30B6 runs

158,2tok/s Ø

Min 66,5Max 233,4

Qwen-AgentWorld-35B-A3B35B78 runs

322,2tok/s Ø

Min 21,8Max 1.481,2

Qwen2.5-0.5B-Instruct3 runs

16,9tok/s Ø

Min 11,7Max 19,6

Qwen2.5-1.5B-Instruct3 runs

6,1tok/s Ø

Min 4,1Max 9,1

Qwen2.5-32B-Instruct-AWQ32B63 runs

127,7tok/s Ø

Min 4,3Max 510,3

Qwen2.5-3B-Instruct3B21 runs

179,2tok/s Ø

Min 0,1Max 364,1

Qwen2.5-72B-Instruct72B32 runs

43,9tok/s Ø

Min 0,8Max 239,2

Qwen2.5-7B-Instruct7B12 runs

179,7tok/s Ø

Min 56,9Max 442,0

Qwen2.5-Coder-32B-Instruct32B104 runs

95,9tok/s Ø

Min 1,5Max 500,6

Qwen2.5-Coder-7B-Instruct3 runs

162,1tok/s Ø

Min 81,1Max 256,7

Qwen3-0.6B0.6B20 runs

248,7tok/s Ø

Min 3,2Max 561,4

Qwen3-1.7B1.7B17 runs

222,2tok/s Ø

Min 3,0Max 492,7

Qwen3-14B14B10 runs

122,9tok/s Ø

Min 10,6Max 280,8

Qwen3-30B-A3B30B3 runs

145,1tok/s Ø

Min 75,9Max 227,0

Qwen3-30B-A3B-Instruct-250730B131 runs

293,9tok/s Ø

Min 1,0Max 1.671,0

Qwen3-30B-A3B-Thinking-250730B115 runs

317,5tok/s Ø

Min 0,6Max 1.683,8

Qwen3-32B32B3 runs

70,0tok/s Ø

Min 28,4Max 117,2

Qwen3-4B4B17 runs

152,5tok/s Ø

Min 0,0Max 325,2

Qwen3-8B8.2B12 runs

140,3tok/s Ø

Min 1,6Max 315,2

Qwen3-Coder-30B-A3B-Instruct30B129 runs

304,2tok/s Ø

Min 1,0Max 1.716,6

Qwen3-Coder-Next57 runs

282,9tok/s Ø

Min 18,1Max 1.253,3

Qwen3-Omni-30B-A3B-Thinking30B74 runs

366,3tok/s Ø

Min 26,2Max 1.673,6

Qwen3-VL-30B-A3B-Instruct30B105 runs

313,9tok/s Ø

Min 16,3Max 1.655,9

Qwen3.5-122B-A10B122B58 runs

86,9tok/s Ø

Min 3,4Max 564,6

Qwen3.5-122B-A10B-Q8122B0 runs

No performance data yet

Qwen3.5-35B-A3B35B111 runs

315,7tok/s Ø

Min 5,6Max 1.520,4

Qwen3.6-27B27B75 runs

130,7tok/s Ø

Min 8,8Max 546,9

Qwen3.6-35B-A3B35B166 runs

248,7tok/s Ø

Min 4,3Max 1.489,1

Qwen3.6-35B-A3B-AWQ35.00B4 runs

No performance data yet

Qwen3.8-27B27B57 runs

101,0tok/s Ø

Min 6,9Max 426,5

Qwen3.8-27B-Uncensored27.00B3 runs

No performance data yet

Qwen3.8-Flash-Next180B56 runs

241,4tok/s Ø

Min 9,0Max 1.564,6

Qwen3.8-Flash-Next-Uncensored-NVFP4180B1 runs

No performance data yet