AI Research

    Qwen3 30B A3B: separate Instruct-2507 and Thinking-2507 editions shipped in July 2025; this profile remains about the original hybrid checkpoint

    This profile covers the original hybrid checkpoint, not the July 2025 editions.

    Separate Instruct-2507 and Thinking-2507 editions shipped in July 2025; this profile remains about the original hybrid checkpoint — verified by the OnAirToday AI Progress Tracker against Alibaba's published documentation.

    O

    OnAirToday

    Newsroom

    2 min
    OnAirToday

    The OnAirToday AI Progress Tracker recorded a verified change to Alibaba's Qwen3 30B A3B (version Qwen3-30B-A3B): Separate Instruct-2507 and Thinking-2507 editions shipped in July 2025; this profile remains about the original hybrid checkpoint.

    Qwen3-30B-A3B is Alibaba's original open-weight, text-only mixture-of-experts model, with 30.5 billion total parameters and 3.3 billion activated parameters, combining thinking and non-thinking modes in one checkpoint. Separate Instruct-2507 and Thinking-2507 editions were released in July 2025; their capabilities and benchmark results should not be attributed to this original model.

    What changed in this review:

    • The 70.9% AIME result is specifically AIME 2025, not OTIS Mock AIME 2024–25.
    • The 57.4% LiveCodeBench result is identified as v5.
    • Hosted pricing is provider-specific rather than an intrinsic price of the open-weight checkpoint.
    • Capability score 68 → 65
    • Flaws and limitations updated

    Teams building on Qwen3 30B A3B 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.

    References:

    • https://huggingface.co/Qwen/Qwen3-30B-A3B
    • https://arxiv.org/abs/2505.09388
    • https://huggingface.co/Qwen/Qwen3-30B-A3B-Instruct-2507
    • https://huggingface.co/Qwen/Qwen3-30B-A3B-Thinking-2507
    AI Analysis

    Generated insight by OnAirToday AI

    The July 2025 editions distinguish newer Instruct and Thinking releases from the original hybrid checkpoint. That distinction helps readers understand which Qwen3 30B A3B version this profile covers.

    Powered by AI • Not a substitute for human analysisMore Insights →
    Read the full story at OnAirToday →
    Share
    Filed under:AI Research

    Comments

    Keep comments relevant and respectful. No links, advertising, self-promotion, or spam.

    0/1000