Model comparison
GPT-4.1 nano vs Llama-3.3-70B-Instruct
Llama-3.3-70B-Instruct is the stronger model overall, scoring 30.6 to 27.9 on the Noometry Index.
Last verified . 25 shared benchmarks.
MetaSummary
- They share 25 benchmarks with published results for both. GPT-4.1 nano scores higher in 3 categories and Llama-3.3-70B-Instruct in 6 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-4.1 nano leads 26.9 to 15.3.
- The biggest single-benchmark swing is MATH Level 5: 70% for GPT-4.1 nano and 41.6% for Llama-3.3-70B-Instruct.
- Llama-3.3-70B-Instruct is cheaper at $0.10 / $0.32 per million input/output tokens, against $0.10 / $0.40 for GPT-4.1 nano.
- GPT-4.1 nano accepts more context: 1.05M tokens versus 128K.
- Llama-3.3-70B-Instruct has downloadable open weights; the other is API-only.
Side by side
| GPT-4.1 nano | Llama-3.3-70B-Instruct | |
|---|---|---|
| Provider | OpenAI | Meta |
| Noometry Index | 27.9 | 30.6 |
| Released | 2025-04-14 | 2024-12-06 |
| Weights | Proprietary | Open |
| Context window | 1.05M | 128K |
| Max output | 33K | 4K |
| Input $ / M tokens | $0.10 | $0.10 |
| Output $ / M tokens | $0.40 | $0.32 |
| Results tracked | 38 | 43 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Llama-3.3-70B-Instruct leads
GPT-4.1 nano: 24.1 (#330), Llama-3.3-70B-Instruct: 31.0 (#290)
| Benchmark | GPT-4.1 nano | Llama-3.3-70B-Instruct |
|---|---|---|
| SciCode | 25.9% | 26% |
| WeirdML | 19% | 14.4% |
| LMArena Coding | 1306 | 1268 |
| Aider Polyglot | 8.9% | — |
| BigCodeBench Instruct | — | 46.9% |
| LiveBench Coding | — | 36.6% |
| BigCodeBench Complete | — | 57.5% |
Agentic & Tool Use Too close to call
GPT-4.1 nano: 26.5 (#104), Llama-3.3-70B-Instruct: 25.8 (#105)
| Benchmark | GPT-4.1 nano | Llama-3.3-70B-Instruct |
|---|---|---|
| Berkeley Function Calling Leaderboard | 33% | 31.9% |
| BALROG | — | 23% |
Reasoning Llama-3.3-70B-Instruct leads
GPT-4.1 nano: 8.5 (#349), Llama-3.3-70B-Instruct: 14.1 (#327)
| Benchmark | GPT-4.1 nano | Llama-3.3-70B-Instruct |
|---|---|---|
| CritPt | 0% | 0% |
| LMArena Hard Prompts | 1286 | 1257 |
| DTBench | 52.5% | 59.5% |
| LMCA | 5.5% | 17.5% |
| Epoch Capabilities Index | 129.62 | 127.33 |
| ARC-AGI-2 | 0% | — |
| SimpleBench | — | 19.9% |
| Kagi LLM Benchmark | 33.3% | — |
| ARC-AGI-1 | 0% | — |
| LiveBench Reasoning | — | 50.8% |
| LiveBench Data Analysis | — | 49.5% |
| ForecastBench | — | 58.6 |
| LiveBench | — | 50.2% |
Math GPT-4.1 nano leads
GPT-4.1 nano: 26.9 (#252), Llama-3.3-70B-Instruct: 15.3 (#298)
| Benchmark | GPT-4.1 nano | Llama-3.3-70B-Instruct |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 28.9% | 5.1% |
| LMArena Math | 1274 | 1267 |
| MATH Level 5 | 70% | 41.6% |
| Omni-MATH | 36.7% | — |
| LiveBench Math | — | 42.2% |
| FrontierMath (Feb 2025 set) | 1% | — |
Knowledge Llama-3.3-70B-Instruct leads
GPT-4.1 nano: 21.8 (#273), Llama-3.3-70B-Instruct: 30.6 (#226)
| Benchmark | GPT-4.1 nano | Llama-3.3-70B-Instruct |
|---|---|---|
| GPQA Diamond | 48.9% | 47.4% |
| LMArena Expert | 1272 | 1225 |
| SimpleQA Verified | 6% | — |
| MMLU-Pro | 55% | — |
| Confabulations | — | 22.8% |
| Vectara Hallucination Rate | — | 4.1% |
| GPQA (HELM) | 50.7% | — |
| MMLU | — | 86.3% |
Multimodal Not comparable
GPT-4.1 nano: 29.2 (#113), Llama-3.3-70B-Instruct: —
| Benchmark | GPT-4.1 nano | Llama-3.3-70B-Instruct |
|---|---|---|
| LMArena Vision | 1063 | — |
Multilingual GPT-4.1 nano leads
GPT-4.1 nano: 41.6 (#205), Llama-3.3-70B-Instruct: 39.9 (#220)
| Benchmark | GPT-4.1 nano | Llama-3.3-70B-Instruct |
|---|---|---|
| LMArena Non-English | 1260 | 1236 |
| LMArena Chinese | 1270 | 1217 |
| LMArena German | 1288 | 1251 |
| LMArena Japanese | 1198 | 1150 |
| LMArena Russian | 1261 | 1252 |
| LMArena French | — | 1281 |
| LMArena Korean | — | 1143 |
| LMArena Spanish | — | 1270 |
Instruction Following Llama-3.3-70B-Instruct leads
GPT-4.1 nano: 67.8 (#193), Llama-3.3-70B-Instruct: 71.1 (#157)
| Benchmark | GPT-4.1 nano | Llama-3.3-70B-Instruct |
|---|---|---|
| LMArena Instruction Following | 1267 | 1242 |
| LiveBench Instruction Following | — | 82.7% |
| IFEval | 84.3% | — |
Long Context Llama-3.3-70B-Instruct leads
GPT-4.1 nano: 23.7 (#296), Llama-3.3-70B-Instruct: 26.4 (#295)
| Benchmark | GPT-4.1 nano | Llama-3.3-70B-Instruct |
|---|---|---|
| Fiction.LiveBench | 25% | 33.3% |
| LMArena Longer Query | 1283 | 1256 |
Writing & Preference Llama-3.3-70B-Instruct leads
GPT-4.1 nano: 40.5 (#243), Llama-3.3-70B-Instruct: 47.6 (#207)
| Benchmark | GPT-4.1 nano | Llama-3.3-70B-Instruct |
|---|---|---|
| LMArena Text | 1285 | 1274 |
| LMArena Creative Writing | 1260 | 1250 |
| LMArena Multi-Turn | 1277 | 1280 |
| EQ-Bench Creative Writing | 946 | — |
| WildBench | 81.2% | — |
| LiveBench Language | — | 39.2% |
Frequently asked questions
Is GPT-4.1 nano better than Llama-3.3-70B-Instruct?
Llama-3.3-70B-Instruct is the stronger model overall, scoring 30.6 to 27.9 on the Noometry Index.
Which is cheaper, GPT-4.1 nano or Llama-3.3-70B-Instruct?
Llama-3.3-70B-Instruct is cheaper. It lists at $0.10 per million input tokens and $0.32 per million output tokens; GPT-4.1 nano lists at $0.10 and $0.40.
Is GPT-4.1 nano or Llama-3.3-70B-Instruct better for coding?
Llama-3.3-70B-Instruct scores higher on coding benchmarks: 31.0 versus 24.1 in the Noometry coding category.
Which has the bigger context window?
GPT-4.1 nano does, with 1.05M tokens against 128K.
How many benchmarks do GPT-4.1 nano and Llama-3.3-70B-Instruct share?
25 benchmarks have published results for both models. GPT-4.1 nano has 38 scored results on Noometry and Llama-3.3-70B-Instruct has 43.