← Back to Directory

Qwen: Qwen3 235B A22B Thinking 2507

QWEN Developer Architecture Profile

OpenRouter ↗HuggingFace ↗
Intelligence (ELO)1046Chatbot Arena Verified
Max Context262,144Tokens
API Cost / 1M$0.20Blended Prompt + Completion

Model Capabilities

    Qwen3-235B-A22B-Thinking-2507 is a high-performance, open-weight Mixture-of-Experts (MoE) language model optimized for complex reasoning tasks. It activates 22B of its 235B parameters per forward pass and natively supports up to 262,144...

    Granular Pricing Matrix

    Input Tokens (Prompt)$0.10 / 1M
    Output Tokens (Completion)$0.10 / 1M

    Pricing data via OpenRouter. Sync: 6/14/2026

    Evaluate Competitors

    VS Engine MatchupQwen: Qwen3 235B A22B Thinking 2507 vs Google: Lyria 3 Clip PreviewVS Engine MatchupQwen: Qwen3 235B A22B Thinking 2507 vs Mistral: Mistral Small 3.2 24BVS Engine MatchupQwen: Qwen3 235B A22B Thinking 2507 vs Google: Gemma 3 4BVS Engine MatchupQwen: Qwen3 235B A22B Thinking 2507 vs NVIDIA: Nemotron 3 Ultra (free)VS Engine MatchupQwen: Qwen3 235B A22B Thinking 2507 vs EssentialAI: Rnj 1 InstructVS Engine MatchupQwen: Qwen3 235B A22B Thinking 2507 vs Body Builder (beta)