Skip to contentT/TTokens or Towers

Compare

LeaderboardEvery tracked model, ranked and priced side by sideBenchmarksWhat each test measures and who leads itPricesAPI, media, and hardware rates with a cost calculator

Plan

Stack plannerSix questions to a priced local and hosted stackShortlistOur picks for subscriptions, APIs, and hardwareEnterpriseSeat plans against self-hosting for a whole team

Build

HardwareGPUs and workstations, and which models they fitHostsWhere to rent open-weight inference, and for how muchHarnessesCoding agents and the models they drive

Lab

FindingsResults from our own evaluationsScanAudit a repo’s AI usage in the browserMethodWhere the numbers come from and how we check them
Data checked September 25, 2026
Data checked September 25, 2026
Plan a stack
T/TTokens or Towers

An independent guide to AI models, what they cost, and the hardware to run them. No account, no API key, no sponsors.

Compare

  • Leaderboard
  • Benchmarks
  • Prices

Plan

  • Stack planner
  • Shortlist
  • Enterprise

Build

  • Hardware
  • Hosts
  • Harnesses

Lab

  • Findings
  • Scan
  • Method
Data checked September 25, 2026 · Every number links to its sourceRSS · Source on GitHub
Leaderboard/MBZUAI
MBZUAI

MBZUAI · K2 Horizon

K2 Horizon MoVA 36B A4B

Open weightsGenerally availableNewApache-2.0

Mid-size 36B/4B-active member of the fully open K2 Horizon family with 512K context.

AA Intelligence
25.3#52 of 142
Price
Self-host
Output speed
—
Context window
524K tokens

Composite indexes

  • Artificial Analysis Intelligence Index#52 of 142
    25independentIntelligence Index v4.3; reasoningArtificial Analysis: K2 Horizon MoVA 36B A4B ↗

    Leader: Claude Opus 5.5 at 58

Coding

  • SciCode#71 of 94
    40%independentreasoningArtificial Analysis: K2 Horizon MoVA 36B A4B ↗

    Leader: Claude Opus 5.5 at 66.9%

  • Terminal-Bench 4.0#86 of 88
    0%independentreasoningArtificial Analysis: K2 Horizon MoVA 36B A4B ↗

    Leader: Claude Mythos 5.1 at 60.9%

  • Terminal-Bench 2.x#52 of 103
    56.6%independentTerminal-Bench 2.1; reasoningArtificial Analysis: K2 Horizon MoVA 36B A4B ↗

    Leader: Claude Fable 5.1 at 91.4%

Agents and tool use

  • GDPval-AA#41 of 103
    1181independentreasoningArtificial Analysis: K2 Horizon MoVA 36B A4B ↗

    Leader: Claude Opus 5.5 at 1846

Reasoning and knowledge

  • GPQA Diamond#66 of 142
    82.2%independentreasoningArtificial Analysis: K2 Horizon MoVA 36B A4B ↗

    Leader: GPT-6 Astra at 96.1%

  • Humanity’s Last Exam#65 of 147
    23.4%independentno tools; reasoningArtificial Analysis: K2 Horizon MoVA 36B A4B ↗

    Leader: Claude Opus 5.5 at 61.4%

  • CritPt#62 of 139
    2.3%independentreasoningArtificial Analysis: K2 Horizon MoVA 36B A4B ↗

    Leader: GPT-5.6 Sol at 32.3%

Vision and long context

  • AA Long Context Reasoning#71 of 139
    71.7%independentreasoningArtificial Analysis: K2 Horizon MoVA 36B A4B ↗

    Leader: Kimi K3 at 88.7%

Other published results

ResultScoreReported bySource
tau2-Bench Banking (AA) · reasoning32%independentArtificial Analysis: K2 Horizon MoVA 36B A4B ↗
AA-Omniscience Index · reasoning-10.9 indexindependentArtificial Analysis: K2 Horizon MoVA 36B A4B ↗

Specs

Released
September 3, 2026
Parameters
36B total, 4B active
Size class
10 to 40B
Input
text
Output
text
Context
524K tokens
Max output
Not published

    Will it fit your budget?

    The planner prices this model against local hardware and other APIs for your workload.

    Plan a stack →

    Compare with

    • TencentTencent Hy325.3
    • XiaomiMiMo-V2.525.2
    • AlibabaQwen3.7-Plus25
    • Thinking Machines LabInkling25