Qwen

Qwen3 235B A22B Instruct 2507

Launched 2025-07-21 · Knowledge up to 2025-06-30 · Listed on OpenRouter 2025-07-21

Prices updated . Benchmarks updated .

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,...

You send / 1M

$0.140

tokens in

It replies / 1M

$0.800

tokens out

Context

262K

tokens it can read at once

Typical request

$0.0006

1,000 words in, 400 out

Benchmarks

BenchmarkScoreKindDateSource
ARC-AGI-21.3%Independent2025-07-25ARC Prize, republished by Epoch AI
CritPt0.0%Independent2025-07-25CritPt, republished by Epoch AI
Arena Elo (text)1,419Independent2026-10-28LMArena

Data privacy

No training or retention: Yes. At least one provider serves it with zero data retention.

ProviderYou send / 1MIt replies / 1MUptimePrivacy policy
DeepInfra · fp8$0.072$0.440n/aRead
DeepInfra · fp8$0.090$0.55079%Read
Nebius · fp8$0.200$0.60097%Read
Venice · fp8$0.150$0.750100%Read
Parasail · fp8$0.140$0.800100%Read
DeepInfra · fp8$0.135$0.82597%Read

Based on OpenRouter and Vercel AI Gateway data. Confirm terms with the provider before sending sensitive data.

What it can do

  • Reads: text
  • Writes: text
  • Weights: Open weights
  • Reasoning: No
  • Tool calling: Yes
  • Longest reply: 16K tokens
  • Repeated text is billed at $0.050 per 1M (prompt caching)

Compare with

Appears in

Estimate your monthly cost with Qwen3 235B A22B Instruct 2507

Sources and licenses

  • OpenRouter: Prices, context windows, modalities and the model list come from the OpenRouter public API. Data: OpenRouter.
  • Vercel AI Gateway: The no-training flag comes from the Vercel AI Gateway model list.
  • models.dev: Lab launch dates, open-weights flags and knowledge cutoffs come from models.dev. Copyright (c) 2025 models.dev, MIT License.
  • Epoch AI: Independent benchmark scores and the Epoch Capabilities Index (ECI) come from Epoch AI's Benchmarking Hub, used under CC BY 4.0. Epoch AI, Capabilities & benchmarking, epoch.ai/benchmarks. Scores Epoch marks as external keep their original license, and the source is named on each score.
  • LMArena: Arena Elo (text, webdev and vision) comes from the LMArena leaderboard dataset (lmarena-ai/leaderboard-dataset), CC BY 4.0, leaderboard date 2026-10-28.
  • Lab-reported scores, when an editor has added them, are the lab's own numbers. Each one links to the source and is labeled lab-reported. They are not used to rank the lists.
  • What a score measures: the benchmark guide.