Model

Amazon Nova Lite

CLOSED

Amazon · Amazon Nova family · released Dec 3, 2024

Low-cost multimodal (text/image/video-in) Nova tier tuned for speed; launched alongside Micro and Pro at re:Invent 2024.

ReasoningCodingVisionFunction callingTool useAgentic

1039.0

Elo · rank #206

Parameters

Undisclosed

Active params

—

Context

300K tokens

Architecture

Transformer, architecture/size undisclosed

License

Proprietary (AWS Bedrock service terms)

Languages

200+

API price (in/out)

$0.06 / $0.24

Modalities

text · vision · video

Benchmark results

Bar shows position within the tracked field; marker = field best

ARC-ChallengeReasoning92.4%#19

best: Llama 3.1 405B · 96.9%

BIG-Bench HardReasoning82.4%#25

best: ERNIE 4.5 300B-A47B · 94.3%

BFCL v3Agents66.6%#18

best: Hunyuan-A13B · 78.3%

ChartQAVision86.8%#16

best: MiniMax-VL-01 · 91.7%

DocVQAVision92.4%#27

best: Qwen2-VL-72B · 96.5%

DROPReasoning80.2#25

best: Hunyuan-T1 · 93.1

GPQA DiamondReasoning42.0%#209

best: GPT-5.6 · 94.6%

GSM8KMath94.5%#13

best: Llama 3.1 405B · 96.8%

HumanEvalCoding85.4%#40

best: Claude Opus 4.5 · 99.4%

IFEvalReasoning89.7%#18

best: Gemma 4 26B A4B · 98.5%

MATH-500Math73.3%#84

best: GPT-5 · 99.4%

MMLUKnowledge80.5%#85

best: OpenAI o3 · 92.9%

MMMUVision56.2%#82

best: Claude Fable 5 · 89.3%

MT-BenchHuman preference8#46

best: Hunyuan-Large (A52B) · 9.4

TextVQAVision80.2%#12

best: Molmo 2 8B · 85.7%

Run it locally

Closed weights — available via API only. No local deployment.

Input / M tok

$0.06

Output / M tok

$0.24

Amazon Nova family

Elo progression across releases

Amazon Nova LiteDec 20241039.0 Amazon Nova MicroDec 2024923.2 Amazon Nova ProDec 20241216.3 Amazon Nova PremierApr 20251235.4

Resources

DOCSdocs.aws.amazon.com/nova/latest/userguide/what-is-nova.html APIAvailable via Amazon API

API price $0.06/$0.24 · each benchmark row carries its own source badge (see methodology)