Llama 3.2 3B Instruct vs Nova Micro: Specs & Benchmark Comparison

Llama 3.2 3B Instruct is developed by Meta, while Nova Micro comes from Amazon. Llama 3.2 3B Instruct was released in September 2024, and Nova Micro followed 2 months later in November 2024. Llama 3.2 3B Instruct has a published size of about 3 billion parameters; Amazon has not disclosed the parameter count of Nova Micro.

The two models share 6 published benchmarks. Nova Micro leads on 6 of them. The widest gaps are on MATH, where Nova Micro scores 69.3% against 48.0%; GSM8k, where Nova Micro scores 92.3% against 77.7%. Averaged across everything we track, Llama 3.2 3B Instruct sits at 55.6% and Nova Micro at 67.0%.

Llama 3.2 3B Instruct is the cheaper API at $0.01 per million input tokens and $0.02 per million output tokens, roughly 3 times cheaper than Nova Micro at $0.03 and $0.14. Both accept a context window of about 128K tokens.

CharacteristicLlama 3.2 3B InstructNova Micro
CompanyMetaAmazon
Release DateSeptember 25, 2024November 20, 2024
Parameters3B—
MultimodalNoNo
Context (input)128K128K
Context (output)128K128K
Input Price / 1M$0.01$0.03
Output Price / 1M$0.02$0.14
Average Score55.6%67.0%
Benchmarks
MATH48.0%69.3%
GSM8k77.7%92.3%
MMLU63.4%77.6%
ARC-C78.6%90.2%
IFEval77.4%87.2%
GPQA32.8%40.0%

Visual Benchmark Comparison

Llama 3.2 3B Instruct
Nova Micro
MATH0.5 vs 0.7
0.5
0.7
GSM8k0.8 vs 0.9
0.8
0.9
MMLU0.6 vs 0.8
0.6
0.8
ARC-C0.8 vs 0.9
0.8
0.9
IFEval0.8 vs 0.9
0.8
0.9
GPQA0.3 vs 0.4
0.3
0.4

Verdict

Both models show equal results — the choice depends on your specific use case.

Overall Performance

Both models show comparable average scores: Llama 3.2 3B Instruct — 0.6, Nova Micro — 0.7.

API Cost

Llama 3.2 3B Instruct is 5.7x cheaper: input $0.01/1M vs $0.03/1M tokens.

Context Window

Same context size: 128K tokens.

Recency

Nova Micro is newer: released 11/20/2024 vs 9/25/2024.

More About These Models

Frequently Asked Questions

Which is better for coding — Llama 3.2 3B Instruct or Nova Micro?
Direct comparison on the SWE-Bench benchmark is not available. We recommend reviewing other metrics on the comparison page.
Which model is cheaper — Llama 3.2 3B Instruct or Nova Micro?
Llama 3.2 3B Instruct is cheaper for input: $0.01 per 1M tokens vs $0.03.
Which has a larger context window — Llama 3.2 3B Instruct or Nova Micro?
Llama 3.2 3B Instruct supports a larger context: 128,000 tokens vs 128,000.

The Llama 3.2 3B Instruct and Nova Micro comparison is updated for 2026. Data includes benchmark results, API pricing, context window size and other specifications. For more detailed information, visit the Llama 3.2 3B Instruct or Nova Micro page. See also the complete list of AI model comparisons.