Model comparison
Llama 3.2 3B vs o3
o3 is the stronger model overall, scoring 47.4 to 28.4 on the Noometry Index. Llama 3.2 3B costs 29× less per token, which makes it the better buy when o3's lead doesn't matter for your workload.
Last verified . 15 shared benchmarks.
MetaSummary
- They share 15 benchmarks with published results for both. Llama 3.2 3B scores higher in 0 categories and o3 in 9 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where o3 leads 63.7 to 24.9.
- The biggest single-benchmark swing is Berkeley Function Calling Leaderboard: 21.9% for Llama 3.2 3B and 63% for o3.
- Llama 3.2 3B is cheaper at $0.05 / $0.33 per million input/output tokens, against $2 / $8 for o3.
- o3 accepts more context: 200K tokens versus 131K.
- Llama 3.2 3B has downloadable open weights; the other is API-only.
Side by side
| Llama 3.2 3B | o3 | |
|---|---|---|
| Provider | Meta | OpenAI |
| Noometry Index | 28.4 | 47.4 |
| Released | 2024-09-24 | 2025-04-16 |
| Weights | Open | Proprietary |
| Context window | 131K | 200K |
| Max output | 118K | 100K |
| Input $ / M tokens | $0.05 | $2 |
| Output $ / M tokens | $0.33 | $8 |
| Results tracked | 18 | 63 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding o3 leads
Llama 3.2 3B: 27.1 (#314), o3: 47.3 (#57)
| Benchmark | Llama 3.2 3B | o3 |
|---|---|---|
| LMArena Coding | 1098 | 1408 |
| SWE-bench Verified | — | 62.3% |
| SWE-bench Verified (bash only) | — | 58.4% |
| Aider Polyglot | — | 81.3% |
| GSO | — | 8.8% |
| WeirdML | — | 52.4% |
| BigCodeBench Instruct | 23.4% | — |
| BigCodeBench Complete | 28.3% | — |
| CadEval | — | 74% |
| ALE-Bench | — | 933.55 |
Agentic & Tool Use o3 leads
Llama 3.2 3B: 19.9 (#135), o3: 34.3 (#42)
| Benchmark | Llama 3.2 3B | o3 |
|---|---|---|
| Berkeley Function Calling Leaderboard | 21.9% | 63% |
| GDPval | — | 30.8% |
| DeepResearch Bench | — | 45.2% |
| OSWorld | — | 23% |
| BALROG | 10.1% | — |
| LMArena Search | — | 1144 |
| METR Time Horizons | — | 65.4% |
Reasoning o3 leads
Llama 3.2 3B: 20.1 (#225), o3: 32.1 (#66)
| Benchmark | Llama 3.2 3B | o3 |
|---|---|---|
| LMArena Hard Prompts | 1095 | 1402 |
| ARC-AGI-2 | — | 6.5% |
| SimpleBench | — | 53.1% |
| Kagi LLM Benchmark | — | 67.6% |
| ARC-AGI-1 | — | 60.8% |
| CritPt | — | 1.4% |
| Chess Puzzles | — | 38% |
| EnigmaEval | — | 13.1% |
| Mystery Game Puzzles | — | 29% |
| DTBench | — | 84.8% |
| LMCA | — | 39.7% |
| Epoch Capabilities Index | — | 146.88 |
| ForecastBench | — | 62.5 |
Math o3 leads
Llama 3.2 3B: 29.9 (#213), o3: 48.9 (#49)
| Benchmark | Llama 3.2 3B | o3 |
|---|---|---|
| LMArena Math | 1126 | 1426 |
| FrontierMath (Tiers 1-3) | — | 33.3% |
| OTIS Mock AIME 2024-2025 | — | 84.4% |
| Omni-MATH | — | 71.4% |
| MATH Level 5 | — | 97.8% |
| FrontierMath (Feb 2025 set) | — | 18.7% |
| FrontierMath Tier 4 (v1) | — | 2.1% |
Knowledge o3 leads
Llama 3.2 3B: 29.1 (#232), o3: 54.5 (#47)
| Benchmark | Llama 3.2 3B | o3 |
|---|---|---|
| LMArena Expert | 1090 | 1403 |
| GPQA Diamond | — | 81.8% |
| Humanity's Last Exam | — | 20.3% |
| SimpleQA Verified | — | 49.4% |
| MMLU-Pro | — | 85.9% |
| Confabulations | — | 14.4% |
| GPQA (HELM) | — | 75.3% |
Multimodal Not comparable
Llama 3.2 3B: —, o3: 40.7 (#35)
| Benchmark | Llama 3.2 3B | o3 |
|---|---|---|
| LMArena Vision | — | 1214 |
| GeoBench | — | 74% |
| VPCT | — | 52% |
Multilingual o3 leads
Llama 3.2 3B: 26.4 (#276), o3: 51.9 (#99)
| Benchmark | Llama 3.2 3B | o3 |
|---|---|---|
| LMArena Non-English | 1019 | 1402 |
| LMArena Chinese | 1017 | 1438 |
| LMArena German | 1056 | 1420 |
| LMArena Russian | 949 | 1407 |
| LMArena French | — | 1430 |
| LMArena Japanese | — | 1404 |
| LMArena Korean | — | 1371 |
| LMArena Spanish | — | 1395 |
Instruction Following o3 leads
Llama 3.2 3B: 56.3 (#270), o3: 73.0 (#122)
| Benchmark | Llama 3.2 3B | o3 |
|---|---|---|
| LMArena Instruction Following | 1089 | 1368 |
| IFEval | — | 86.9% |
Long Context o3 leads
Llama 3.2 3B: 33.4 (#256), o3: 53.2 (#6)
| Benchmark | Llama 3.2 3B | o3 |
|---|---|---|
| LMArena Longer Query | 1101 | 1372 |
| Fiction.LiveBench | — | 88.9% |
| CL-bench | — | 17.8% |
Writing & Preference o3 leads
Llama 3.2 3B: 24.9 (#302), o3: 63.7 (#62)
| Benchmark | Llama 3.2 3B | o3 |
|---|---|---|
| LMArena Text | 1110 | 1410 |
| LMArena Creative Writing | 1094 | 1359 |
| EQ-Bench Creative Writing | 595 | 1676 |
| LMArena Multi-Turn | 1105 | 1405 |
| Short-Story Creative Writing | — | 83.9% |
| WildBench | — | 86.1% |
Frequently asked questions
Is Llama 3.2 3B better than o3?
o3 is the stronger model overall, scoring 47.4 to 28.4 on the Noometry Index. Llama 3.2 3B costs 29× less per token, which makes it the better buy when o3's lead doesn't matter for your workload.
Which is cheaper, Llama 3.2 3B or o3?
Llama 3.2 3B is cheaper. It lists at $0.05 per million input tokens and $0.33 per million output tokens; o3 lists at $2 and $8.
Is Llama 3.2 3B or o3 better for coding?
o3 scores higher on coding benchmarks: 47.3 versus 27.1 in the Noometry coding category.
Which has the bigger context window?
o3 does, with 200K tokens against 131K.
How many benchmarks do Llama 3.2 3B and o3 share?
15 benchmarks have published results for both models. Llama 3.2 3B has 18 scored results on Noometry and o3 has 63.