Model comparison
GPT-5.4 nano vs Llama 3.2 1B
GPT-5.4 nano is the stronger model overall, scoring 41.9 to 20.1 on the Noometry Index. Llama 3.2 1B costs 6.6× less per token, which makes it the better buy when GPT-5.4 nano's lead doesn't matter for your workload.
Last verified . 17 shared benchmarks.
MetaSummary
- They share 17 benchmarks with published results for both. GPT-5.4 nano scores higher in 8 categories and Llama 3.2 1B in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where GPT-5.4 nano leads 41.9 to 7.2.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 87.8% for GPT-5.4 nano and 0.6% for Llama 3.2 1B.
- Llama 3.2 1B is cheaper at $0.027 / $0.20 per million input/output tokens, against $0.20 / $1.25 for GPT-5.4 nano.
- GPT-5.4 nano accepts more context: 400K tokens versus 60K.
- Llama 3.2 1B has downloadable open weights; the other is API-only.
Side by side
| GPT-5.4 nano | Llama 3.2 1B | |
|---|---|---|
| Provider | OpenAI | Meta |
| Noometry Index | 41.9 | 20.1 |
| Released | 2026-03-17 | 2024-09-24 |
| Weights | Proprietary | Open |
| Context window | 400K | 60K |
| Max output | 128K | 54K |
| Input $ / M tokens | $0.20 | $0.027 |
| Output $ / M tokens | $1.25 | $0.20 |
| Results tracked | 40 | 22 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5.4 nano leads
GPT-5.4 nano: 43.6 (#84), Llama 3.2 1B: 21.1 (#338)
| Benchmark | GPT-5.4 nano | Llama 3.2 1B |
|---|---|---|
| LMArena Coding | 1405 | 1070 |
| SciCode | 46.9% | — |
| WeirdML | 49.2% | — |
| BigCodeBench Instruct | — | 8.2% |
| BigCodeBench Complete | — | 11.3% |
| ALE-Bench | 1,005 | — |
Agentic & Tool Use Not comparable
GPT-5.4 nano: —, Llama 3.2 1B: 14.6 (#150)
| Benchmark | GPT-5.4 nano | Llama 3.2 1B |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 10.8% |
| BALROG | — | 6.6% |
Reasoning GPT-5.4 nano leads
GPT-5.4 nano: 23.7 (#173), Llama 3.2 1B: 16.2 (#308)
| Benchmark | GPT-5.4 nano | Llama 3.2 1B |
|---|---|---|
| Chess Puzzles | 30% | 0% |
| LMArena Hard Prompts | 1381 | 1044 |
| Epoch Capabilities Index | 145.81 | 101.99 |
| ARC-AGI-2 | 5.7% | — |
| Kagi LLM Benchmark | 39.7% | — |
| ARC-AGI-1 | 51.5% | — |
| CritPt | 9.3% | — |
| Mystery Game Puzzles | 9% | — |
| DTBench | 80.3% | — |
| LMCA | 36.9% | — |
| ForecastBench | 57.3 | — |
Math GPT-5.4 nano leads
GPT-5.4 nano: 40.9 (#88), Llama 3.2 1B: 10.4 (#313)
| Benchmark | GPT-5.4 nano | Llama 3.2 1B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 87.8% | 0.6% |
| LMArena Math | 1406 | 1086 |
| FrontierMath (Tiers 1-3) | 44.9% | — |
| FrontierMath Tier 4 | 12.2% | — |
| ProofBench | 5% | — |
| FrontierMath (Feb 2025 set) | 25.9% | — |
| FrontierMath Tier 4 (v1) | 6.3% | — |
Knowledge GPT-5.4 nano leads
GPT-5.4 nano: 41.9 (#103), Llama 3.2 1B: 7.2 (#312)
| Benchmark | GPT-5.4 nano | Llama 3.2 1B |
|---|---|---|
| GPQA Diamond | 78.5% | 23.9% |
| LMArena Expert | 1396 | 1007 |
| SimpleQA Verified | 11.7% | — |
| Vectara Hallucination Rate | 3.1% | — |
Multimodal Not comparable
GPT-5.4 nano: 36.7 (#78), Llama 3.2 1B: —
| Benchmark | GPT-5.4 nano | Llama 3.2 1B |
|---|---|---|
| LMArena Vision | 1196 | — |
Multilingual GPT-5.4 nano leads
GPT-5.4 nano: 48.6 (#140), Llama 3.2 1B: 23.8 (#292)
| Benchmark | GPT-5.4 nano | Llama 3.2 1B |
|---|---|---|
| LMArena Non-English | 1359 | 973 |
| LMArena Chinese | 1392 | 959 |
| LMArena German | 1367 | 1014 |
| LMArena Russian | 1363 | 941 |
| LMArena French | 1396 | — |
| LMArena Japanese | 1343 | — |
| LMArena Korean | 1320 | — |
| LMArena Spanish | 1371 | — |
Instruction Following GPT-5.4 nano leads
GPT-5.4 nano: 71.9 (#144), Llama 3.2 1B: 52.4 (#290)
| Benchmark | GPT-5.4 nano | Llama 3.2 1B |
|---|---|---|
| LMArena Instruction Following | 1362 | 1031 |
Long Context GPT-5.4 nano leads
GPT-5.4 nano: 41.6 (#137), Llama 3.2 1B: 31.9 (#274)
| Benchmark | GPT-5.4 nano | Llama 3.2 1B |
|---|---|---|
| LMArena Longer Query | 1366 | 1050 |
Writing & Preference GPT-5.4 nano leads
GPT-5.4 nano: 55.7 (#142), Llama 3.2 1B: 21.3 (#310)
| Benchmark | GPT-5.4 nano | Llama 3.2 1B |
|---|---|---|
| LMArena Text | 1372 | 1055 |
| LMArena Creative Writing | 1314 | 1033 |
| LMArena Multi-Turn | 1382 | 1030 |
| EQ-Bench Creative Writing | — | 200 |
Frequently asked questions
Is GPT-5.4 nano better than Llama 3.2 1B?
GPT-5.4 nano is the stronger model overall, scoring 41.9 to 20.1 on the Noometry Index. Llama 3.2 1B costs 6.6× less per token, which makes it the better buy when GPT-5.4 nano's lead doesn't matter for your workload.
Which is cheaper, GPT-5.4 nano or Llama 3.2 1B?
Llama 3.2 1B is cheaper. It lists at $0.027 per million input tokens and $0.20 per million output tokens; GPT-5.4 nano lists at $0.20 and $1.25.
Is GPT-5.4 nano or Llama 3.2 1B better for coding?
GPT-5.4 nano scores higher on coding benchmarks: 43.6 versus 21.1 in the Noometry coding category.
Which has the bigger context window?
GPT-5.4 nano does, with 400K tokens against 60K.
How many benchmarks do GPT-5.4 nano and Llama 3.2 1B share?
17 benchmarks have published results for both models. GPT-5.4 nano has 40 scored results on Noometry and Llama 3.2 1B has 22.