Qwen: Qwen3 235B A22B Instruct 2507

qwen/qwen3-235b-a22b-2507

Qwen3-235B-A22B-Instruct-2507 is a multilingual, instruction-tuned mixture-of-experts language model based on the Qwen3-235B architecture, with 22B active parameters per forward pass. It is optimized for general-purpose text generation, including instruction following,...

Prompt $/1M

$0.090

Completion $/1M

$0.550

Context

262,144 tok

Best provider

78 tok/s · WandB

Specs

Model id: qwen/qwen3-235b-a22b-2507
Display name: Qwen: Qwen3 235B A22B Instruct 2507
Modality: text->text
Tokenizer: Qwen3
Instruct type: —
Context length: 262,144
Max completion: 16,384
Prompt $/1M: $0.090
Completion $/1M: $0.550
Free tier: no

Capabilities

Function callingno
JSON modeno
System promptno
Visionno
Moderatedno

Benchmark scores

Each axis is one benchmark, normalized to the best model in the field.

No benchmark data for this model.

Speed by provider

Provider	tok/s	TTFT	Total latency
WandBfastest	78	269 ms	324 ms
Alibaba	47	459 ms	524 ms
Cerebras	46	606 ms	908 ms
Google	41	469 ms	724 ms
Nebius	40	544 ms	746 ms
StreamLake	37	710 ms	845 ms
AtlasCloud	35	1,250 ms	1,380 ms
Together	35	552 ms	750 ms
Parasail	27	472 ms	556 ms
Novita	24	740 ms	918 ms
DeepInfra	19	406 ms	588 ms
Venice	17	656 ms	730 ms
Friendli	13	545 ms	790 ms
Chutes	8	1,830 ms	2,013 ms
SiliconFlow	8	2,002 ms	2,387 ms
GMICloud	7	1,054 ms	1,166 ms