gpt·basedElo-ranked, value-sorted
Sign in

Stay updated

Get an email when a cheaper, stronger model than yours appears on the leaderboard. No spam.

gpt·based

The LLM leaderboard that tells you when to switch. Arena rankings joined with OpenRouter pricing, refreshed daily.

Use cases

  • Best for coding
  • Best for multilingual
  • Cheapest worth using
  • Compare two models

Browse

  • Text models
  • WebDev models
  • Vision models
  • Image generation
  • Image edit

Data sources

  • LMArena
  • Providers

Developers

  • API on RapidAPI
  • OpenAPI spec

Contact

  • Contact page
  • [email protected]
© 2026 gptbasedDaily refresh, never paid placement.
    gpt·basedElo-ranked, value-sorted
    Sign in
    Back to Text · German

    Qwen: Qwen3 30B A3B Instruct 2507

    Released Jul 29, 2025
    text→text
    Set alert
    Ranked in:Text
    Try it on:
    OOpenRouter

    Qwen3-30B-A3B-Instruct-2507 is a 30.5B-parameter mixture-of-experts language model from Qwen, with 3.3B active parameters per inference. It operates in non-thinking mode and is designed for high-quality instruction following, multilingual understanding, and...

    Best overall rank
    #150
    ▼ 2 vs 7d ago
    Total arena votes
    142,016
    Across 29 (subset, category) entries
    Providers on OR
    7
    Alibaba + others
    First seen
    2026-05-08
    72 days tracked

    Trend

    Pricing

    Per token

    Input (per 1M)
    $0.100
    Output (per 1M)
    $0.300

    Capabilities

    Modality
    text->text
    Context
    262K tokens
    Tokenizer
    Qwen3
    Knowledge cutoff
    2025-06-30
    Moderation
    Unmoderated
    Hugging Face
    Qwen/Qwen3-30B-A3B-Instruct-2507
    Supported parameters
    frequency_penalty
    logit_bias
    logprobs
    max_tokens
    min_p
    presence_penalty
    repetition_penalty
    response_format
    seed
    stop
    structured_outputs
    temperature
    tool_choice
    tools
    top_k
    top_logprobs
    top_p

    Providers

    ProviderRegion / TagContextInput/MOutput/MUptime 1dLatencyQuantCache
    Alibabaalibaba131K$0.130$0.52099.96%---
    WandBwandb/bf16262K$0.100$0.30099.94%-
    bf16
    -
    StreamLakestreamlake128K$0.048$0.19399.70%---
    Nebiusnebius/fp8262K$0.100$0.30098.99%-
    fp8
    -
    DekaLLMdekallm262K$0.090$0.30097.97%---
    SiliconFlowsiliconflow/fp8262K$0.090$0.30097.26%-
    fp8
    -
    Phalaphala262K$0.150$0.55043.16%---

    Per-category ranks

    Text

    RankCategoryVariantTrendEloVotes
    #115Japanese
    -
    1325704
    #116French
    -
    1418307
    #119Chinese
    -
    14361,324
    #120Spanish
    -
    1387633
    #132Coding
    -
    14394,658
    #132Industry Business And Management And Financial Operations
    -
    13964,172
    #133Non English
    -
    137312,663
    #134Industry Mathematical
    -
    14001,207
    #134Industry Software And It Services
    -
    14298,195
    #137German
    -
    1360487
    #144Korean
    -
    1314567
    #145Hard Prompts
    -
    140711,066
    #147Hard Prompts English
    -
    14165,624
    #149Exclude Ties
    -
    135916,684
    #149Industry Medicine And Healthcare
    -
    13991,347
    #150Overall
    -
    138323,724
    #151Expert
    -
    13921,147
    #151Multi Turn
    -
    13834,073
    #151Polish
    -
    13451,494
    #153English
    -
    139311,050
    #153Russian
    -
    13661,227
    #154Longer Query
    -
    13814,880
    #154Math
    -
    13801,428
    #155Industry Life And Physical And Social Science
    -
    13943,680
    #155Instruction Following
    -
    13675,962
    #157Industry Legal And Government
    -
    13821,492
    #162Industry Entertainment And Sports And Media
    -
    13324,144
    #169Industry Writing And Literature And Language
    -
    13435,095
    #179Creative Writing
    -
    13212,982