Nova Micro

← AI Models
Amazon
2024-12-03
Proprietary
Modality:
Intelligence
10.3
#436/523
Coding
4.1
#390/429
Math
6
#245/265
Speed
302 tok/s
TTFT: 354.00s
Pricing
$35.00 / $0.14
per 1M tokens (in/out)
Google Preferred Source

A text-only model that delivers lowest-latency responses at very low cost while maintaining strong performance on core language tasks. Optimized for speed and efficiency while preserving high accuracy on key benchmarks.

Nova Micro is Amazon’s offering for high-speed token processing. It operates at 301.809 tokens per second and is priced at $35 per million input tokens, targeting professional users.

When to Use Nova Micro

✓ Best For

  • Natural language processing tasks
  • Data analysis and insights generation
  • Real-time applications requiring fast response times

✗ Not Ideal For

  • Users needing advanced coding capabilities
  • Applications requiring high mathematical computation

How Nova Micro Compares

Intelligence Index · Higher is better

GoogleMicrosoftAmazonAnthropic

Benchmark Profile

Coding Index

MetaGoogleAmazonLiquid AI

Output Speed · tok/s

AlibabaGoogleAmazonNVIDIA

Math Index

IBMOpenAIAmazonMistral

Intelligence · Coding · Math

Intelligence Coding Math

All Benchmark Scores (31)

BenchmarkScore
Intelligence Index 10.3
Coding Index 4.1
Math Index 6
MMLU-Pro 531%
GPQA 358%
LiveCodeBench 14%
HLE 47%
SciCode 9.4%
IFBench 29.4%
LCR 9.7%
TerminalBench Hard 1.5%
Tau2 14%
AIME 8%
AIME 2025 6%
MATH 500 70.3%
ARC-C 90.2%
BBH 79.5%
BFCL 56.2%
CRAG 43.1%
DROP 79.3%
FinQA 65.2%
GSM8k 92.3%
HumanEval 81.1%
IFEval 87.2%
MATH 69.3%
MMLU 77.6%
SQuALITY 18.8%
Translation en→Set1 COMET22 88.5%
Translation en→Set1 spBleu 40.2%
Translation Set1→en COMET22 88.7%
Translation Set1→en spBleu 42.6%

Data: Artificial Analysis · Updated: April 9, 2026

Frequently Asked Questions (15)

When was Nova Micro released?
Nova Micro was released on December 3, 2024.
Who created Nova Micro?
Nova Micro was created by Amazon.
How intelligent is Nova Micro?
Nova Micro scores 10 on the Artificial Analysis Intelligence Index, placing it below average among other non-reasoning models in a similar price tier (median: 10).
How fast is Nova Micro?
Nova Micro generates output at 287.5 tokens per second (based on Amazon's API), which is well above average compared to other non-reasoning models in a similar price tier (median: 89.3 t/s).
What is the latency of Nova Micro?
Nova Micro has a time to first token (TTFT) of 0.79s (based on Amazon's API), which is very competitive compared to other non-reasoning models in a similar price tier (median: 0.91s).
How much does Nova Micro cost?
Nova Micro costs $0.04 per 1M input tokens (better than average, median: $0.05) and $0.14 per 1M output tokens (better than average, median: $0.15), based on Amazon's API.
What is Nova Micro API pricing?
Nova Micro costs $0.04 per 1M input tokens and $0.14 per 1M output tokens (based on Amazon's API). For a blended rate (3:1 input to output ratio), this is $0.06 per 1M tokens. Pricing may vary by provider.
How verbose is Nova Micro?
When evaluated on the Intelligence Index, Nova Micro generated 7.1M output tokens, which is better than average compared to other non-reasoning models in a similar price tier (median: 7.7M).
Is Nova Micro a reasoning model?
No, Nova Micro is not a reasoning model. It provides direct responses without extended chain-of-thought reasoning.
What input modalities does Nova Micro support?
Nova Micro supports text only input.
What output modalities does Nova Micro support?
Nova Micro supports text only output.
Can Nova Micro process images?
No, Nova Micro does not support image input. It can only process text.
Is Nova Micro multimodal?
No, Nova Micro is not multimodal. It only supports text only input.
What is the context window of Nova Micro?
Nova Micro has a context window of 130k tokens. This determines how much text and conversation history the model can process in a single request.
Is Nova Micro open source?
No, Nova Micro is proprietary. The model weights are not publicly available.