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/Google
Google

Google · Gemini 3

Gemini 3.5 Flash-Lite

Closed weightsGenerally availableproprietary

Google's current cost-efficient Flash-Lite tier for high-volume agentic tasks, translation and simple data processing.

AA Intelligence
22.2#62 of 142
Input / output per 1M
$0.30/ $2.50
Output speed
350tok/s
Context window
1M tokens

Composite indexes

  • Artificial Analysis Intelligence Index#62 of 142
    22independentv4.3.2, defaultArtificial Analysis: Gemini 3.5 Flash-Lite ↗

    Leader: Claude Opus 5.5 at 58

Coding

  • SciCode#66 of 94
    41.3%independentAA run, defaultArtificial Analysis: Gemini 3.5 Flash-Lite ↗

    Leader: Claude Opus 5.5 at 66.9%

  • Terminal-Bench 4.0#43 of 88
    1%independentAA run (Terminal-Bench 4.0), defaultArtificial Analysis: Gemini 3.5 Flash-Lite ↗

    Leader: Claude Mythos 5.1 at 60.9%

  • Terminal-Bench 2.x#58 of 103
    53.6%independentAA run (Terminal-Bench 2.1), defaultArtificial Analysis: Gemini 3.5 Flash-Lite ↗

    Leader: Claude Fable 5.1 at 91.4%

Agents and tool use

  • GDPval-AA#57 of 103
    970independentGDPval-AA v2.1, defaultArtificial Analysis: Gemini 3.5 Flash-Lite ↗

    Leader: Claude Opus 5.5 at 1846

Reasoning and knowledge

  • GPQA Diamond#64 of 142
    83.8%independentAA run, defaultArtificial Analysis: Gemini 3.5 Flash-Lite ↗

    Leader: GPT-6 Astra at 96.1%

  • Humanity’s Last Exam#75 of 147
    18.8%independentAA run, no tools, defaultArtificial Analysis: Gemini 3.5 Flash-Lite ↗

    Leader: Claude Opus 5.5 at 61.4%

  • CritPt#94 of 139
    0%independentAA run, defaultArtificial Analysis: Gemini 3.5 Flash-Lite ↗

    Leader: GPT-5.6 Sol at 32.3%

  • ARC-AGI-2#21 of 25
    10.3%independenthigh thinking; ARC Prize verifiedARC Prize leaderboard data (semi-private) ↗

    Leader: GPT-6 Astra at 95%

Vision and long context

  • AA Long Context Reasoning#56 of 139
    76%independentAA-LCR v1.1, defaultArtificial Analysis: Gemini 3.5 Flash-Lite ↗

    Leader: Kimi K3 at 88.7%

  • MMMU-Pro#21 of 72
    79%independentAA run, defaultArtificial Analysis: Gemini 3.5 Flash-Lite ↗

    Leader: Claude Opus 5.5 at 87.7%

Human preference

  • LMArena Text#32 of 54
    1456independentgemini-3.5-flash-lite; 26,165 votesLMArena text leaderboard ↗

    Leader: Claude Fable 5 at 1506

  • LMArena WebDev#42 of 54
    1447independentgemini-3.5-flash-lite; 228 votes (preliminary)LMArena Code Arena | WebDev leaderboard ↗

    Leader: Claude Opus 5.5 at 1818

Specs

Released
July 21, 2026
Parameters
Undisclosed
Size class
Undisclosed
Input
text, image, audio, video
Output
text
Context
1M tokens
Max output
Not published
Cached input
$0.03 / 1M
Blended 70/30
$0.96 / 1M
First token
8.77s
  • Price: Gemini Developer API pricing ↗ Single rate for text/image/video/audio input. Batch $0.15/$1.25.
  • Speed: Artificial Analysis: Gemini 3.5 Flash-Lite ↗

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-Pro V221.7
  • NVIDIANemotron 3 Ultra 550B A55B22.9
  • MiniMaxMiniMax-M2.723
  • AlibabaQwen3.6-27B21