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/Ant Group
Ant Group

Ant Group · Ring 2.6

Ring-2.6-1T

Open weightsGenerally availableMIT

Ant Group's open-weight 1T-parameter (63B active) reasoning model.

AA Intelligence
16.6#75 of 142
Price
Self-host
Output speed
122tok/s
Context window
262K tokens

Composite indexes

  • Artificial Analysis Intelligence Index#75 of 142
    17independentv4.3.2Artificial Analysis: Ring-2.6-1T ↗

    Leader: Claude Opus 5.5 at 58

Coding

  • Terminal-Bench 4.0#54 of 88
    0.5%independentAAArtificial Analysis: Ring-2.6-1T ↗

    Leader: Claude Mythos 5.1 at 60.9%

  • Terminal-Bench 2.x#68 of 103
    43.1%independentAA, Terminal-Bench 2.1Artificial Analysis: Ring-2.6-1T ↗

    Leader: Claude Fable 5.1 at 91.4%

  • SciCode#56 of 94
    45%independentAAArtificial Analysis: Ring-2.6-1T ↗

    Leader: Claude Opus 5.5 at 66.9%

Agents and tool use

  • τ²-Bench#17 of 88
    92.4%independentAA, telecomArtificial Analysis: Ring-2.6-1T ↗

    Leader: GLM-5.2 at 99.1%

  • GDPval-AA#67 of 103
    734independentGDPval-AA v2.1Artificial Analysis: Ring-2.6-1T ↗

    Leader: Claude Opus 5.5 at 1846

Reasoning and knowledge

  • Humanity’s Last Exam#70 of 147
    21.6%independentAA, no toolsArtificial Analysis: Ring-2.6-1T ↗

    Leader: Claude Opus 5.5 at 61.4%

  • GPQA Diamond#56 of 142
    85.7%independentAAArtificial Analysis: Ring-2.6-1T ↗

    Leader: GPT-6 Astra at 96.1%

  • CritPt#56 of 139
    3.7%independentAAArtificial Analysis: Ring-2.6-1T ↗

    Leader: GPT-5.6 Sol at 32.3%

Vision and long context

  • AA Long Context Reasoning#74 of 139
    70%independentAAArtificial Analysis: Ring-2.6-1T ↗

    Leader: Kimi K3 at 88.7%

Other published results

ResultScoreReported bySource
Terminal-Bench Hard (AA) · AA28.8%independentArtificial Analysis: Ring-2.6-1T ↗
IFBench (AA) · AA44.6%independentArtificial Analysis: Ring-2.6-1T ↗
AA-Omniscience Index · AA-37.7 index (-100..100)independentArtificial Analysis: Ring-2.6-1T ↗

Specs

Released
May 8, 2026
Parameters
1T total, 63B active
Size class
1 to 2T
Input
text
Output
text
Context
262K tokens
Max output
Not published
First token
3.98s
  • Speed: Artificial Analysis: Ring-2.6-1T (median across providers) ↗

Will it fit your budget?

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

Plan a stack →

Compare with

  • GoogleGemma 4 26B A4B16.7
  • AnthropicClaude Haiku 4.516.9
  • ByteDanceDoubao Seed Code16.9
  • MetaMuse Glimmer17.5