Model comparison
GPT-5.2 vs Llama 2-70B
GPT-5.2 is the stronger model overall, scoring 54.1 to 24.4 on the Noometry Index.
Last verified . 22 shared benchmarks.
MetaSummary
- They share 22 benchmarks with published results for both. GPT-5.2 scores higher in 8 categories and Llama 2-70B in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-5.2 leads 60.0 to 8.1.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 96.1% for GPT-5.2 and 0% for Llama 2-70B.
- Llama 2-70B has downloadable open weights; the other is API-only.
Side by side
| GPT-5.2 | Llama 2-70B | |
|---|---|---|
| Provider | OpenAI | Meta |
| Noometry Index | 54.1 | 24.4 |
| Released | 2025-12-11 | 2023-07-18 |
| Weights | Proprietary | Open |
| Context window | 400K | — |
| Max output | 128K | — |
| Input $ / M tokens | $1.75 | — |
| Output $ / M tokens | $14 | — |
| Results tracked | 67 | 35 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5.2 leads
GPT-5.2: 51.6 (#37), Llama 2-70B: 31.4 (#286)
| Benchmark | GPT-5.2 | Llama 2-70B |
|---|---|---|
| LMArena Coding | 1447 | 1079 |
| SWE-bench Verified | 73.8% | — |
| SWE-bench Verified (bash only) | 72.8% | — |
| LMArena WebDev | 1416 | — |
| SWE-bench Multilingual | 66.7% | — |
| GSO | 27.4% | — |
| WeirdML | 72.2% | — |
| ALE-Bench | 1,294 | — |
| AlgoTune | 2.05 | — |
Agentic & Tool Use Not comparable
GPT-5.2: 40.2 (#24), Llama 2-70B: —
| Benchmark | GPT-5.2 | Llama 2-70B |
|---|---|---|
| Terminal-Bench | 64.9% | — |
| Berkeley Function Calling Leaderboard | 55.9% | — |
| GDPval | 49.7% | — |
| Remote Labor Index | 2.5% | — |
| τ²-bench Airline | 83% | — |
| τ²-bench Banking | 32.2% | — |
| τ²-bench Retail | 81.6% | — |
| τ²-bench Telecom | 89.7% | — |
| DeepResearch Bench | 41.1% | — |
| LMArena Search | 1207 | — |
| METR Time Horizons | 75.3% | — |
| Vending-Bench 2 | 3,591 | — |
Reasoning GPT-5.2 leads
GPT-5.2: 50.2 (#35), Llama 2-70B: 14.4 (#325)
| Benchmark | GPT-5.2 | Llama 2-70B |
|---|---|---|
| LMArena Hard Prompts | 1445 | 1073 |
| DTBench | 90.9% | 41.6% |
| Epoch Capabilities Index | 153.45 | 113.79 |
| ForecastBench | 60.1 | 51.4 |
| ARC-AGI-2 | 52.9% | — |
| SimpleBench | 45.8% | — |
| Kagi LLM Benchmark | 73.3% | — |
| NYT Connections (extended) | 83.6% | — |
| ARC-AGI-1 | 86.2% | — |
| Chess Puzzles | 49% | — |
| EnigmaEval | 10.4% | — |
| EBR-Bench | 23% | — |
| Mystery Game Puzzles | 23% | — |
| LMCA | 43.9% | — |
| BIG-Bench Hard | — | 64.9% |
| CommonsenseQA 2.0 | — | 50% |
| HellaSwag | — | 85.3% |
| LAMBADA | — | 78.9% |
| PIQA | — | 82.8% |
| WinoGrande | — | 80.2% |
Math GPT-5.2 leads
GPT-5.2: 60.0 (#38), Llama 2-70B: 8.1 (#326)
| Benchmark | GPT-5.2 | Llama 2-70B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 96.1% | 0% |
| LMArena Math | 1440 | 1091 |
| FrontierMath (Tiers 1-3) | 67.4% | — |
| FrontierMath Tier 4 | 31.7% | — |
| MathArena Final-Answer Competitions | 72% | — |
| ProofBench | 15% | — |
| MATH Level 5 | — | 3.3% |
| FrontierMath (Feb 2025 set) | 40.7% | — |
| FrontierMath Tier 4 (v1) | 18.8% | — |
| GSM8K | — | 69.6% |
Knowledge GPT-5.2 leads
GPT-5.2: 59.3 (#32), Llama 2-70B: 7.4 (#310)
| Benchmark | GPT-5.2 | Llama 2-70B |
|---|---|---|
| GPQA Diamond | 91.4% | 26.3% |
| LMArena Expert | 1445 | 1039 |
| Humanity's Last Exam | 27.8% | — |
| SimpleQA Verified | 37.1% | — |
| Vectara Hallucination Rate | 8.4% | — |
| ARC (AI2) Challenge | — | 78.3% |
| BoolQ | — | 88.6% |
| MMLU | — | 69.9% |
| OpenBookQA | — | 60.2% |
| TriviaQA | — | 87.6% |
Multimodal Not comparable
GPT-5.2: 51.3 (#7), Llama 2-70B: —
| Benchmark | GPT-5.2 | Llama 2-70B |
|---|---|---|
| LMArena Vision | 1268 | — |
| VPCT | 84% | — |
| Furniture Assembly | 38.3% | — |
| LMArena Document | 1405 | — |
Multilingual GPT-5.2 leads
GPT-5.2: 53.4 (#67), Llama 2-70B: 27.7 (#274)
| Benchmark | GPT-5.2 | Llama 2-70B |
|---|---|---|
| LMArena Non-English | 1425 | 1045 |
| LMArena Chinese | 1460 | 995 |
| LMArena French | 1455 | 1090 |
| LMArena German | 1448 | 1041 |
| LMArena Japanese | 1420 | 927 |
| LMArena Korean | 1392 | 964 |
| LMArena Russian | 1440 | 1083 |
| LMArena Spanish | 1433 | 1143 |
Instruction Following GPT-5.2 leads
GPT-5.2: 74.7 (#89), Llama 2-70B: 54.9 (#278)
| Benchmark | GPT-5.2 | Llama 2-70B |
|---|---|---|
| LMArena Instruction Following | 1417 | 1071 |
Long Context GPT-5.2 leads
GPT-5.2: 44.0 (#78), Llama 2-70B: 32.3 (#270)
| Benchmark | GPT-5.2 | Llama 2-70B |
|---|---|---|
| LMArena Longer Query | 1428 | 1062 |
| CL-bench | 18.2% | — |
Writing & Preference GPT-5.2 leads
GPT-5.2: 66.8 (#32), Llama 2-70B: 32.3 (#279)
| Benchmark | GPT-5.2 | Llama 2-70B |
|---|---|---|
| LMArena Text | 1439 | 1115 |
| LMArena Creative Writing | 1401 | 1075 |
| LMArena Multi-Turn | 1458 | 1088 |
| EQ-Bench Creative Writing | 1703 | — |
Frequently asked questions
Is GPT-5.2 better than Llama 2-70B?
GPT-5.2 is the stronger model overall, scoring 54.1 to 24.4 on the Noometry Index.
Is GPT-5.2 or Llama 2-70B better for coding?
GPT-5.2 scores higher on coding benchmarks: 51.6 versus 31.4 in the Noometry coding category.
How many benchmarks do GPT-5.2 and Llama 2-70B share?
22 benchmarks have published results for both models. GPT-5.2 has 67 scored results on Noometry and Llama 2-70B has 35.