Qwen 3 32B available; 128K context, 100+ languages, ~491 TPS
2 features1 enhancementThis release2 featuresNew capabilities1 enhancementImprovements to existing featuresAI-tallied from the release notes
From the original release noteView original ↗
Added Qwen 3 32B, the latest generation Qwen model, offers advancements in reasoning, instruction-following, agent capabilities, and multilingual support, with seamless switching between thinking and non-thinking modes.
Key Features: 128K context window; support for 100+ languages; tool use and JSON mode; ~491 TPS; $0.29/1M input, $0.59/1M output.
Performance: 93.8% ArenaHard, 81.4% AIME 2024, 65.7% LiveCodeBench, 30.3% BFCL, 73.0% MultiIF, 72.9% AIME 2025, 71.6% LiveBench.
Changed
Python SDK updated to v0.26.0/v0.27.0, TypeScript SDK to v0.23.0/v0.24.0. Added include_images field to search_settings for agentic tooling systems; added code_results (png/text) to executed tool outputs.
Fetched August 11, 2026
