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/Amazon
Amazon

Amazon · Nova

Amazon Nova Micro

Closed weightsGenerally availableproprietary

Text-only, lowest-cost Nova model for high-volume simple tasks.

AA Intelligence
5.9#131 of 142
Input / output per 1M
$0.035/ $0.14
Output speed
265tok/s
Context window
130K tokens

Composite indexes

  • Artificial Analysis Intelligence Index#131 of 142
    6independentIntelligence Index v4.3; non-reasoning; estimatedArtificial Analysis: Nova Micro ↗

    Leader: Claude Opus 5.5 at 58

Coding

  • LiveCodeBench#53 of 56
    14%independentnon-reasoningArtificial Analysis: Nova Micro ↗

    Leader: Fugu at 92.9%

Agents and tool use

  • τ²-Bench#76 of 88
    14%independenttelecom; non-reasoningArtificial Analysis: Nova Micro ↗

    Leader: GLM-5.2 at 99.1%

Reasoning and knowledge

  • GPQA Diamond#136 of 142
    35.8%independentnon-reasoningArtificial Analysis: Nova Micro ↗

    Leader: GPT-6 Astra at 96.1%

  • Humanity’s Last Exam#129 of 147
    4.6%independentno tools; non-reasoningArtificial Analysis: Nova Micro ↗

    Leader: Claude Opus 5.5 at 61.4%

  • CritPt#120 of 139
    0%independentnon-reasoningArtificial Analysis: Nova Micro ↗

    Leader: GPT-5.6 Sol at 32.3%

Vision and long context

  • AA Long Context Reasoning#123 of 139
    13%independentnon-reasoningArtificial Analysis: Nova Micro ↗

    Leader: Kimi K3 at 88.7%

Other published results

ResultScoreReported bySource
aime-2025 · AIME 2025; non-reasoning6%independentArtificial Analysis: Nova Micro ↗
Terminal-Bench Hard (AA) · non-reasoning1.5%independentArtificial Analysis: Nova Micro ↗
IFBench · non-reasoning29.4%independentArtificial Analysis: Nova Micro ↗
AA-Omniscience Index · non-reasoning-48.5 indexindependentArtificial Analysis: Nova Micro ↗

Specs

Released
December 3, 2024
Parameters
Undisclosed
Size class
Undisclosed
Input
text
Output
text
Context
130K tokens
Max output
Not published
Cached input
$0.009 / 1M
Blended 70/30
$0.067 / 1M
First token
0.83s
  • Price: Amazon Bedrock pricing ↗ Bedrock on-demand rate; cache read $0.00875/M.
  • Speed: Artificial Analysis leaderboard (Nova Micro, median across providers or first-party as shown) ↗

Will it fit your budget?

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

Plan a stack →

Compare with

  • MicrosoftPhi-45.9
  • Liquid AILFM2 24B A2B5.9
  • MistralMinistral 3 14B6
  • MicrosoftPhi-4 Multimodal Instruct5.8