Model comparison
Claude Sonnet 4 vs Llama 3.2 3B
Claude Sonnet 4 is the stronger model overall, scoring 40.6 to 28.4 on the Noometry Index. Llama 3.2 3B costs 50× less per token, which makes it the better buy when Claude Sonnet 4's lead doesn't matter for your workload.
Last verified . 14 shared benchmarks.
MetaSummary
- They share 14 benchmarks with published results for both. Claude Sonnet 4 scores higher in 9 categories and Llama 3.2 3B in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where Claude Sonnet 4 leads 57.3 to 24.9.
- Llama 3.2 3B is cheaper at $0.05 / $0.33 per million input/output tokens, against $3 / $15 for Claude Sonnet 4.
- Claude Sonnet 4 accepts more context: 200K tokens versus 131K.
- Llama 3.2 3B has downloadable open weights; the other is API-only.
Side by side
| Claude Sonnet 4 | Llama 3.2 3B | |
|---|---|---|
| Provider | Anthropic | Meta |
| Noometry Index | 40.6 | 28.4 |
| Released | 2025-05-22 | 2024-09-24 |
| Weights | Proprietary | Open |
| Context window | 200K | 131K |
| Max output | 64K | 118K |
| Input $ / M tokens | $3 | $0.05 |
| Output $ / M tokens | $15 | $0.33 |
| Results tracked | 58 | 18 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Claude Sonnet 4 leads
Claude Sonnet 4: 44.0 (#70), Llama 3.2 3B: 27.1 (#314)
| Benchmark | Claude Sonnet 4 | Llama 3.2 3B |
|---|---|---|
| LMArena Coding | 1414 | 1098 |
| SWE-bench Verified (bash only) | 64.9% | — |
| Aider Polyglot | 61.3% | — |
| SciCode | 40% | — |
| GSO | 4.9% | — |
| WeirdML | 46.1% | — |
| BigCodeBench Instruct | — | 23.4% |
| BigCodeBench Complete | — | 28.3% |
| ALE-Bench | 655.35 | — |
Agentic & Tool Use Claude Sonnet 4 leads
Claude Sonnet 4: 38.3 (#29), Llama 3.2 3B: 19.9 (#135)
| Benchmark | Claude Sonnet 4 | Llama 3.2 3B |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 21.9% |
| TheAgentCompany | 33.1% | — |
| Cybench | 35% | — |
| DeepResearch Bench | 46.6% | — |
| OSWorld | 43.9% | — |
| BALROG | — | 10.1% |
| METR Time Horizons | 62% | — |
Reasoning Claude Sonnet 4 leads
Claude Sonnet 4: 23.0 (#169), Llama 3.2 3B: 20.1 (#225)
| Benchmark | Claude Sonnet 4 | Llama 3.2 3B |
|---|---|---|
| LMArena Hard Prompts | 1372 | 1095 |
| ARC-AGI-2 | 5.9% | — |
| SimpleBench | 45.5% | — |
| Kagi LLM Benchmark | 73% | — |
| ARC-AGI-1 | 40% | — |
| CritPt | 0.3% | — |
| EnigmaEval | 3.1% | — |
| DTBench | 77.1% | — |
| LMCA | 29% | — |
| Epoch Capabilities Index | 141.69 | — |
| ForecastBench | 60.2 | — |
Math Claude Sonnet 4 leads
Claude Sonnet 4: 41.5 (#71), Llama 3.2 3B: 29.9 (#213)
| Benchmark | Claude Sonnet 4 | Llama 3.2 3B |
|---|---|---|
| LMArena Math | 1375 | 1126 |
| OTIS Mock AIME 2024-2025 | 71.1% | — |
| Omni-MATH | 60.2% | — |
| MATH Level 5 | 84.4% | — |
| FrontierMath (Feb 2025 set) | 4.1% | — |
| FrontierMath Tier 4 (v1) | 0% | — |
Knowledge Claude Sonnet 4 leads
Claude Sonnet 4: 41.6 (#99), Llama 3.2 3B: 29.1 (#232)
| Benchmark | Claude Sonnet 4 | Llama 3.2 3B |
|---|---|---|
| LMArena Expert | 1371 | 1090 |
| GPQA Diamond | 79.2% | — |
| Humanity's Last Exam | 7.8% | — |
| MMLU-Pro | 84.3% | — |
| Confabulations | 13.2% | — |
| Vectara Hallucination Rate | 10.3% | — |
| GPQA (HELM) | 70.6% | — |
Multimodal Not comparable
Claude Sonnet 4: 25.6 (#116), Llama 3.2 3B: —
| Benchmark | Claude Sonnet 4 | Llama 3.2 3B |
|---|---|---|
| LMArena Vision | 1191 | — |
| GeoBench | 37% | — |
| VPCT | 34% | — |
| MindCube | 44.8% | — |
Multilingual Claude Sonnet 4 leads
Claude Sonnet 4: 46.9 (#151), Llama 3.2 3B: 26.4 (#276)
| Benchmark | Claude Sonnet 4 | Llama 3.2 3B |
|---|---|---|
| LMArena Non-English | 1333 | 1019 |
| LMArena Chinese | 1351 | 1017 |
| LMArena German | 1331 | 1056 |
| LMArena Russian | 1355 | 949 |
| LMArena French | 1364 | — |
| LMArena Japanese | 1301 | — |
| LMArena Korean | 1293 | — |
| LMArena Spanish | 1356 | — |
Instruction Following Claude Sonnet 4 leads
Claude Sonnet 4: 71.8 (#141), Llama 3.2 3B: 56.3 (#270)
| Benchmark | Claude Sonnet 4 | Llama 3.2 3B |
|---|---|---|
| LMArena Instruction Following | 1375 | 1089 |
| IFEval | 84% | — |
Long Context Too close to call
Claude Sonnet 4: 33.6 (#254), Llama 3.2 3B: 33.4 (#256)
| Benchmark | Claude Sonnet 4 | Llama 3.2 3B |
|---|---|---|
| LMArena Longer Query | 1398 | 1101 |
| Fiction.LiveBench | 46.9% | — |
Writing & Preference Claude Sonnet 4 leads
Claude Sonnet 4: 57.3 (#127), Llama 3.2 3B: 24.9 (#302)
| Benchmark | Claude Sonnet 4 | Llama 3.2 3B |
|---|---|---|
| LMArena Text | 1351 | 1110 |
| LMArena Creative Writing | 1345 | 1094 |
| EQ-Bench Creative Writing | 1483 | 595 |
| LMArena Multi-Turn | 1376 | 1105 |
| Short-Story Creative Writing | 81.4% | — |
| WildBench | 83.8% | — |
Frequently asked questions
Is Claude Sonnet 4 better than Llama 3.2 3B?
Claude Sonnet 4 is the stronger model overall, scoring 40.6 to 28.4 on the Noometry Index. Llama 3.2 3B costs 50× less per token, which makes it the better buy when Claude Sonnet 4's lead doesn't matter for your workload.
Which is cheaper, Claude Sonnet 4 or Llama 3.2 3B?
Llama 3.2 3B is cheaper. It lists at $0.05 per million input tokens and $0.33 per million output tokens; Claude Sonnet 4 lists at $3 and $15.
Is Claude Sonnet 4 or Llama 3.2 3B better for coding?
Claude Sonnet 4 scores higher on coding benchmarks: 44.0 versus 27.1 in the Noometry coding category.
Which has the bigger context window?
Claude Sonnet 4 does, with 200K tokens against 131K.
How many benchmarks do Claude Sonnet 4 and Llama 3.2 3B share?
14 benchmarks have published results for both models. Claude Sonnet 4 has 58 scored results on Noometry and Llama 3.2 3B has 18.