Nova Micro 1.0 vs Llama 3.2 3B Instruct
Updated October 21, 2026. 33 benchmarks have a result for both. Coding is ranked on the Vibe Coding Index; every other board is shown as published.
Decision reading
Neither model has a coding score yet, so the two cannot be ranked on coding.
Nova Micro 1.0 has the edge where output tokens cost 2.4x less, and where what matters is τ²-Bench Telecom or Humanity's Last Exam (Artificial Analysis run), where it leads.
- Shared results
- 33
- Nova Micro 1.0 only
- 6
- Llama 3.2 3B Instruct only
- 1
- Like-for-like categories
- 7 of 9
Shared results can support a head-to-head reading. Results only one model has describe coverage, not superiority.
Which one for your work
Each card is decided by one number both models carry. Where only one has it, the card says so and makes no call.
Coding
Vibe Coding Index, six public coding benchmarks
No call
Neither has a coding score yet.
One side not measured
Intelligence
AA Intelligence Index
Nova Micro 1.0
Nova Micro 1.0 leads, 5.9 against 5.7.
Same board, both measured
Long documents
Listed context window
Llama 3.2 3B Instruct
Llama 3.2 3B Instruct takes 131K tokens against 128K.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
Nova Micro 1.0
Nova Micro 1.0 $0.0001, Llama 3.2 3B Instruct $0.0002 per turn.
List prices
At a glance
- Coding index
- Nova Micro 1.0 not benchmarked, Llama 3.2 3B Instruct not benchmarked.
- Widest category
- Knowledge, 17.7 points apart on average over 5 shared benchmarks.
- Widest single gap
- LMArena Text: Chinese: Nova Micro 1.0 1209 against Llama 3.2 3B Instruct 1017, 32.0 points apart on a 0 to 100 scale.
- Output price
- Nova Micro 1.0 $0.14 against Llama 3.2 3B Instruct $0.3325 per million output tokens, 2.4x apart.
- Context window
- Nova Micro 1.0 takes 128K tokens, Llama 3.2 3B Instruct 131K.
Benchmark by benchmark
First the score: boards each model leads, per category, where both have a result. Then every board, a category to a fold; bars are each result on a 0 to 100 scale, so compare the two in a row, not one row with another.
| Category | Nova Micro 1.0 | Bars | Llama 3.2 3B Instruct | Ties |
|---|---|---|---|---|
| Coding | 3 | 0 | ||
| Intelligence | 1 | 0 | ||
| Agentic work | 0 | 1 | ||
| Reasoning | 4 | 1 | ||
| Math | 4 | 0 | ||
| Knowledge | 5 | 0 | ||
| General | 14 | 0 | ||
| All shared boards | 31 | 2 |
CodingLike-for-likeSix of these boards feed the Vibe Coding Index30
| Benchmark | Nova Micro 1.0 | Llama 3.2 3B Instruct | Leader |
|---|---|---|---|
| LiveCodeBench (Artificial Analysis run) | 14.0% | 8.3% | Nova Micro 1.0 |
| Terminal-Bench Hard (Artificial Analysis run) | 1.5% | Not measured | Only Nova Micro 1.0 measured |
| LMArena Text: coding | 1218 | 1098 | Nova Micro 1.0 |
| LMArena Text: software and IT | 1215 | 1103 | Nova Micro 1.0 |
| Neither model has a result on the other 34 boards here | |||
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index10
| Benchmark | Nova Micro 1.0 | Llama 3.2 3B Instruct | Leader |
|---|---|---|---|
| AA Intelligence Index | 5.9 | 5.7 | Nova Micro 1.0 |
| Neither model has a result on the other 1 board here | |||
Agentic workLike-for-likePublic boards, shown as published, kept out of the index01
| Benchmark | Nova Micro 1.0 | Llama 3.2 3B Instruct | Leader |
|---|---|---|---|
| τ²-Bench Telecom | 14.0% | 21.1% | Llama 3.2 3B Instruct |
| BALROG | Not measured | 10.1% | Only Llama 3.2 3B Instruct measured |
| Neither model has a result on the other 37 boards here | |||
ReasoningLike-for-likePublic boards, shown as published, kept out of the index41
| Benchmark | Nova Micro 1.0 | Llama 3.2 3B Instruct | Leader |
|---|---|---|---|
| GPQA Diamond (Artificial Analysis run) | 35.8% | 25.5% | Nova Micro 1.0 |
| Humanity's Last Exam (Artificial Analysis run) | 4.6% | 5.3% | Llama 3.2 3B Instruct |
| LMArena Text: hard prompts | 1191 | 1095 | Nova Micro 1.0 |
| LMArena Text: expert | 1183 | 1090 | Nova Micro 1.0 |
| LMArena Text: hard prompts in English | 1203 | 1141 | Nova Micro 1.0 |
| Neither model has a result on the other 16 boards here | |||
MathLike-for-likePublic boards, shown as published, kept out of the index40
| Benchmark | Nova Micro 1.0 | Llama 3.2 3B Instruct | Leader |
|---|---|---|---|
| AA Math Index | 6.0 | 3.3 | Nova Micro 1.0 |
| AIME 2025 (Artificial Analysis run) | 6.0% | 3.3% | Nova Micro 1.0 |
| LMArena Text: math | 1206 | 1126 | Nova Micro 1.0 |
| LMArena Text: mathematical work | 1212 | 1128 | Nova Micro 1.0 |
| Neither model has a result on the other 9 boards here | |||
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index50
| Benchmark | Nova Micro 1.0 | Llama 3.2 3B Instruct | Leader |
|---|---|---|---|
| MMLU Pro (Artificial Analysis run) | 53.1% | 34.7% | Nova Micro 1.0 |
| MMLU | 70.8% | Not measured | Only Nova Micro 1.0 measured |
| LMArena Text: science | 1210 | 1106 | Nova Micro 1.0 |
| LMArena Text: medicine and healthcare | 1207 | 1104 | Nova Micro 1.0 |
| LMArena Text: legal and government | 1221 | 1115 | Nova Micro 1.0 |
| LMArena Text: business and finance | 1182 | 1074 | Nova Micro 1.0 |
| Neither model has a result on the other 11 boards here | |||
Web designNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | Nova Micro 1.0 | Llama 3.2 3B Instruct | Leader |
|---|---|---|---|
| Neither model has a result on 9 boards here | |||
GeneralLike-for-likePublic boards, shown as published, kept out of the index140
| Benchmark | Nova Micro 1.0 | Llama 3.2 3B Instruct | Leader |
|---|---|---|---|
| IFBench | 29.4% | 26.2% | Nova Micro 1.0 |
| Long context reasoning (AA-LCR) | 13.0% | 4.3% | Nova Micro 1.0 |
| LMArena Text | 1208 | 1110 | Nova Micro 1.0 |
| LMArena Text: instruction following | 1174 | 1089 | Nova Micro 1.0 |
| LMArena Text: creative writing | 1172 | 1094 | Nova Micro 1.0 |
| LMArena Text: English | 1224 | 1180 | Nova Micro 1.0 |
| LMArena Text: non-English | 1186 | 1019 | Nova Micro 1.0 |
| LMArena Text: Chinese | 1209 | 1017 | Nova Micro 1.0 |
| LMArena Text: German | 1192 | 1056 | Nova Micro 1.0 |
| LMArena Text: Japanese | 1154 | Not measured | Only Nova Micro 1.0 measured |
| LMArena Text: Korean | 1150 | Not measured | Only Nova Micro 1.0 measured |
| LMArena Text: French | 1238 | Not measured | Only Nova Micro 1.0 measured |
| LMArena Text: Spanish | 1225 | Not measured | Only Nova Micro 1.0 measured |
| LMArena Text: Russian | 1185 | 949 | Nova Micro 1.0 |
| LMArena Text: multi-turn | 1178 | 1105 | Nova Micro 1.0 |
| LMArena Text: longer queries | 1205 | 1100 | Nova Micro 1.0 |
| LMArena Text: writing and language | 1192 | 1096 | Nova Micro 1.0 |
| LMArena Text: entertainment and media | 1163 | 1083 | Nova Micro 1.0 |
| Neither model has a result on the other 28 boards here | |||
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | Nova Micro 1.0 | Llama 3.2 3B Instruct | Leader |
|---|---|---|---|
| Neither model has a result on 5 boards here | |||
What each workload costs
Three fixed token mixes at each model's listed price, so per-token rates become a bill.
Chat turn
1K input + 500 output tokens
Nova Micro 1.0 costs less
Repository review
50K input + 3K output tokens
Nova Micro 1.0 costs less
Agent loop
200K input + 10K output tokens
Nova Micro 1.0 costs less
Listed OpenRouter rates, without cache or batch discounts. A workload that does not fit a model's context window would need splitting.
Specification differences
The leading value in a row is marked; a row with a missing value names no leader. Our two rows are marked in red.
| Specification | Amazon Nova Micro 1.0 | Meta Llama 3.2 3B Instruct |
|---|---|---|
| Maker | Amazon | Meta Llama |
| Weights | No public weights listed | Open weights |
| Released | December 5, 2024 | September 25, 2024 |
| Context window | 128K | 131K |
| Input $/Mtok | $0.035 | $0.0505 |
| Output $/Mtok | $0.14 | $0.3325 |
| Coding index | Not benchmarked | Not benchmarked |
| Independent measurements behind it | 0 | 0 |
| Cost per solved task | Not computable | Not computable |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is Nova Micro 1.0 better than Llama 3.2 3B Instruct for coding?
Neither model has a coding score yet, so the two cannot be ranked on coding. Nova Micro 1.0 has the edge where output tokens cost 2.4x less, and where what matters is τ²-Bench Telecom or Humanity's Last Exam (Artificial Analysis run), where it leads.
Which is cheaper, Nova Micro 1.0 or Llama 3.2 3B Instruct?
Nova Micro 1.0 costs $0.035 per million input tokens and $0.14 per million output tokens; Llama 3.2 3B Instruct costs $0.0505 and $0.3325. Cost per solved task is known for neither of them.
Which has the bigger context window, Nova Micro 1.0 or Llama 3.2 3B Instruct?
Llama 3.2 3B Instruct, with 131K tokens against 128K.
How many benchmarks is the Nova Micro 1.0 vs Llama 3.2 3B Instruct comparison based on?
33 benchmarks have a result for both models as of October 21, 2026. Nova Micro 1.0's coding index rests on 0 independent benchmarks and Llama 3.2 3B Instruct's on 0. Where only one model has a result, the page says so instead of naming a winner.