Model comparison
Llama 3.2 3B vs Llama 4 Maverick
Llama 4 Maverick is the stronger model overall, scoring 30.9 to 28.9 on the Noometry Index. Llama 3.2 3B costs 2.5× less per token, which makes it the better buy when Llama 4 Maverick's lead doesn't matter for your workload.
Last verified . 17 shared benchmarks.
MetaSummary
- They share 17 benchmarks with published results for both. Llama 3.2 3B scores higher in 4 categories and Llama 4 Maverick in 5 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in multilingual, where Llama 4 Maverick leads 42.2 to 26.2.
- The biggest single-benchmark swing is BigCodeBench Complete: 28.3% for Llama 3.2 3B and 61.4% for Llama 4 Maverick.
- Llama 3.2 3B is cheaper at $0.05 / $0.33 per million input/output tokens, against $0.19 / $0.65 for Llama 4 Maverick.
- Llama 3.2 3B accepts more context: 131K tokens versus 128K.
Side by side
| Llama 3.2 3B | Llama 4 Maverick | |
|---|---|---|
| Provider | Meta | Meta |
| Noometry Index | 28.9 | 30.9 |
| Released | 2024-09-24 | 2025-04-05 |
| Weights | Open | Open |
| Context window | 131K | 128K |
| Max output | 118K | 4K |
| Input $ / M tokens | $0.05 | $0.19 |
| Output $ / M tokens | $0.33 | $0.65 |
| Results tracked | 18 | 54 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Llama 3.2 3B leads
Llama 3.2 3B: 27.6 (#319), Llama 4 Maverick: 26.6 (#324)
| Benchmark | Llama 3.2 3B | Llama 4 Maverick |
|---|---|---|
| BigCodeBench Instruct | 23.4% | 49.7% |
| LMArena Coding | 1098 | 1302 |
| BigCodeBench Complete | 28.3% | 61.4% |
| SWE-bench Verified (bash only) | — | 21% |
| Aider Polyglot | — | 15.6% |
| SciCode | — | 33.1% |
| WeirdML | — | 24.5% |
| ALE-Bench | — | 172.97 |
Agentic & Tool Use Llama 4 Maverick leads
Llama 3.2 3B: 20.1 (#143), Llama 4 Maverick: 28.2 (#91)
| Benchmark | Llama 3.2 3B | Llama 4 Maverick |
|---|---|---|
| Berkeley Function Calling Leaderboard | 21.9% | 37.3% |
| BALROG | 10.1% | — |
Reasoning Llama 3.2 3B leads
Llama 3.2 3B: 21.0 (#228), Llama 4 Maverick: 10.1 (#342)
| Benchmark | Llama 3.2 3B | Llama 4 Maverick |
|---|---|---|
| LMArena Hard Prompts | 1095 | 1281 |
| ARC-AGI-2 | — | 0% |
| SimpleBench | — | 27.7% |
| Kagi LLM Benchmark | — | 55.9% |
| NYT Connections (extended) | — | 8% |
| ARC-AGI-1 | — | 4.4% |
| CritPt | — | 0% |
| EnigmaEval | — | 0.6% |
| DTBench | — | 61.9% |
| LMCA | — | 15.9% |
| Epoch Capabilities Index | — | 132.2 |
| ForecastBench | — | 57.5 |
Math Llama 3.2 3B leads
Llama 3.2 3B: 32.4 (#214), Llama 4 Maverick: 26.0 (#262)
| Benchmark | Llama 3.2 3B | Llama 4 Maverick |
|---|---|---|
| LMArena Math | 1126 | 1299 |
| OTIS Mock AIME 2024-2025 | — | 20.6% |
| Omni-MATH | — | 42.2% |
| MATH Level 5 | — | 73% |
| FrontierMath (Feb 2025 set) | — | 0.7% |
Knowledge Llama 4 Maverick leads
Llama 3.2 3B: 29.7 (#235), Llama 4 Maverick: 33.4 (#204)
| Benchmark | Llama 3.2 3B | Llama 4 Maverick |
|---|---|---|
| LMArena Expert | 1090 | 1259 |
| GPQA Diamond | — | 67% |
| Humanity's Last Exam | — | 5.7% |
| MMLU-Pro | — | 81% |
| Confabulations | — | 22.6% |
| Vectara Hallucination Rate | — | 8.2% |
| GPQA (HELM) | — | 65% |
Multimodal Not comparable
Llama 3.2 3B: —, Llama 4 Maverick: 31.6 (#105)
| Benchmark | Llama 3.2 3B | Llama 4 Maverick |
|---|---|---|
| LMArena Vision | — | 1142 |
| GeoBench | — | 52% |
| SpatialViz-Bench | — | 31.8% |
Multilingual Llama 4 Maverick leads
Llama 3.2 3B: 26.2 (#281), Llama 4 Maverick: 42.2 (#195)
| Benchmark | Llama 3.2 3B | Llama 4 Maverick |
|---|---|---|
| LMArena Non-English | 1019 | 1269 |
| LMArena Chinese | 1017 | 1277 |
| LMArena German | 1056 | 1291 |
| LMArena Russian | 949 | 1286 |
| LMArena French | — | 1259 |
| LMArena Japanese | — | 1207 |
| LMArena Korean | — | 1203 |
| LMArena Spanish | — | 1293 |
Instruction Following Llama 4 Maverick leads
Llama 3.2 3B: 56.0 (#275), Llama 4 Maverick: 71.7 (#146)
| Benchmark | Llama 3.2 3B | Llama 4 Maverick |
|---|---|---|
| LMArena Instruction Following | 1089 | 1267 |
| IFEval | — | 90.8% |
Long Context Llama 3.2 3B leads
Llama 3.2 3B: 33.4 (#261), Llama 4 Maverick: 31.4 (#279)
| Benchmark | Llama 3.2 3B | Llama 4 Maverick |
|---|---|---|
| LMArena Longer Query | 1100 | 1280 |
| Fiction.LiveBench | — | 46.2% |
Writing & Preference Llama 4 Maverick leads
Llama 3.2 3B: 24.7 (#307), Llama 4 Maverick: 38.8 (#252)
| Benchmark | Llama 3.2 3B | Llama 4 Maverick |
|---|---|---|
| LMArena Text | 1110 | 1287 |
| LMArena Creative Writing | 1094 | 1267 |
| EQ-Bench Creative Writing | 595 | 860 |
| LMArena Multi-Turn | 1105 | 1289 |
| Short-Story Creative Writing | — | 62% |
| WildBench | — | 80% |
Frequently asked questions
Is Llama 3.2 3B better than Llama 4 Maverick?
Llama 4 Maverick is the stronger model overall, scoring 30.9 to 28.9 on the Noometry Index. Llama 3.2 3B costs 2.5× less per token, which makes it the better buy when Llama 4 Maverick's lead doesn't matter for your workload.
Which is cheaper, Llama 3.2 3B or Llama 4 Maverick?
Llama 3.2 3B is cheaper. It lists at $0.05 per million input tokens and $0.33 per million output tokens; Llama 4 Maverick lists at $0.19 and $0.65.
Is Llama 3.2 3B or Llama 4 Maverick better for coding?
Llama 3.2 3B scores higher on coding benchmarks: 27.6 versus 26.6 in the Noometry coding category.
Which has the bigger context window?
Llama 3.2 3B does, with 131K tokens against 128K.
How many benchmarks do Llama 3.2 3B and Llama 4 Maverick share?
17 benchmarks have published results for both models. Llama 3.2 3B has 18 scored results on Noometry and Llama 4 Maverick has 54.