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/Kuaishou
Kuaishou

Kuaishou · KAT-Coder V2.5

KAT-Coder-Pro V2.5

Closed weightsGenerally availableproprietary

Kuaishou's proprietary agentic coding model served via StreamLake, tuned for long-horizon repository tasks in Claude Code-style harnesses.

AA Intelligence
—
Input / output per 1M
$0.74/ $2.96
Output speed
—
Context window
262K tokens

Coding

  • SWE-bench Pro#7 of 31
    65.2%vendorClaude Code harnessarXiv 2607.05471: KAT-Coder-V2.5 Technical Report (Table 4) ↗

    Leader: Claude Opus 5.5 at 89.9%

  • Terminal-Bench 2.x#49 of 103
    60.7%vendorTerminal-Bench 2.1arXiv 2607.05471: KAT-Coder-V2.5 Technical Report (Table 4) ↗

    Leader: Claude Fable 5.1 at 91.4%

  • SciCode#40 of 94
    50.3%vendorarXiv 2607.05471: KAT-Coder-V2.5 Technical Report (Table 4) ↗

    Leader: Claude Opus 5.5 at 66.9%

Other published results

ResultScoreReported bySource
PinchBench (avg)94.9%vendorarXiv 2607.05471: KAT-Coder-V2.5 Technical Report (Table 4) ↗
KAT Code Bench53.1%vendorarXiv 2607.05471: KAT-Coder-V2.5 Technical Report (Table 4) ↗
KAT Claw Bench85.5%vendorarXiv 2607.05471: KAT-Coder-V2.5 Technical Report (Table 4) ↗

Specs

Released
July 10, 2026
Parameters
Undisclosed
Size class
Undisclosed
Input
text
Output
text
Context
262K tokens
Max output
Not published
Cached input
$0.15 / 1M
Blended 70/30
$1.406 / 1M
  • Price: OpenRouter endpoints API: kwaipilot/kat-coder-pro-v2.5 (AtlasCloud) ↗ NOT first-party: only OpenRouter provider is AtlasCloud. First-party StreamLake (Kuaishou) pay-as-you-go price page could not be fetched; third-party blogs report the same $0.74/$2.96/$0.15. Release date = OpenRouter listing date.

Will it fit your budget?

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

Plan a stack →

Compare with

  • KuaishouKAT-Coder-V2.5-Dev
  • KuaishouKAT-Coder-Pro V221.7
  • OpenAIGPT-6 Astra52.7
  • OpenAIGPT-6 Sol47.5