Qwen3 30b A3b Instruct: the 61.3% AIME result is identified specifically as AIME 2025

The OnAirToday AI Progress Tracker recorded a verified change to Alibaba's Qwen3 30b A3b Instruct (version Qwen3-30B-A3B-Instruct-2507): The 61.3% AIME result is identified specifically as AIME 2025.

Alibaba's Qwen3-30B-A3B-Instruct-2507 is an open-weight, text-only mixture-of-experts model designed for non-thinking instruction following. It supports a documented 262,144-token context window and multilingual assistance, coding and tool-use workflows, but is distinct from both the original hybrid Qwen3-30B-A3B and the separate Thinking-2507 edition.

What changed in this review:

• Unconfirmed OTIS Mock AIME and LMArena Text figures are not included in the documented benchmark results.
• Capability score 63 → 64
• Flaws and limitations updated

Teams building on Qwen3 30b A3b Instruct should confirm limits, pricing and benchmark coverage against the vendor's own documentation before relying on these figures.

The full profile — capabilities, benchmark citations, known flaws and risk guidance — is on the tracker at /ai-models/qwen3-30b-a3b-instruct.

References:

• https://huggingface.co/Qwen/Qwen3-30B-A3B-Instruct-2507
• https://github.com/QwenLM/Qwen3