Model comparison
DeepSeek-R1 vs Llama 2-70B
DeepSeek-R1 is the stronger model overall, scoring 42.0 to 24.1 on the Noometry Index.
Last verified . 22 shared benchmarks.
DeepSeek
MetaSummary
- They share 22 benchmarks with published results for both. DeepSeek-R1 scores higher in 8 categories and Llama 2-70B in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where DeepSeek-R1 leads 44.3 to 7.4.
- The biggest single-benchmark swing is MATH Level 5: 96.6% for DeepSeek-R1 and 3.3% for Llama 2-70B.
- Llama 2-70B has downloadable open weights; the other is API-only.
Side by side
| DeepSeek-R1 | Llama 2-70B | |
|---|---|---|
| Provider | DeepSeek | Meta |
| Noometry Index | 42.0 | 24.1 |
| Released | 2025-01-20 | 2023-07-18 |
| Weights | Proprietary | Open |
| Context window | 164K | — |
| Max output | 64K | — |
| Input $ / M tokens | $0.50 | — |
| Output $ / M tokens | $2.15 | — |
| Results tracked | 52 | 35 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek-R1 leads
DeepSeek-R1: 45.9 (#63), Llama 2-70B: 30.8 (#281)
| Benchmark | DeepSeek-R1 | Llama 2-70B |
|---|---|---|
| LMArena Coding | 1427 | 1079 |
| Aider Polyglot | 71.4% | — |
| SciCode | 35.7% | — |
| WeirdML | 41.6% | — |
| LiveBench Coding | 66.7% | — |
| ALE-Bench | 804.12 | — |
| AlgoTune | 1.7 | — |
Agentic & Tool Use Not comparable
DeepSeek-R1: 30.4 (#68), Llama 2-70B: —
| Benchmark | DeepSeek-R1 | Llama 2-70B |
|---|---|---|
| DeepResearch Bench | 35.1% | — |
| BALROG | 34.9% | — |
| METR Time Horizons | 53.8% | — |
Reasoning DeepSeek-R1 leads
DeepSeek-R1: 18.7 (#253), Llama 2-70B: 13.7 (#324)
| Benchmark | DeepSeek-R1 | Llama 2-70B |
|---|---|---|
| LMArena Hard Prompts | 1415 | 1073 |
| Epoch Capabilities Index | 141.3 | 113.59 |
| ForecastBench | 60 | 51.4 |
| ARC-AGI-2 | 1.3% | — |
| SimpleBench | 40.8% | — |
| Kagi LLM Benchmark | 69.4% | — |
| ARC-AGI-1 | 21.2% | — |
| CritPt | 1.1% | — |
| LiveBench Reasoning | 83.2% | — |
| DTBench | — | 41.6% |
| LiveBench Data Analysis | 69.8% | — |
| BIG-Bench Hard | — | 64.9% |
| CommonsenseQA 2.0 | — | 50% |
| HellaSwag | — | 85.3% |
| LAMBADA | — | 78.9% |
| LiveBench | 71.6% | — |
| PIQA | — | 82.8% |
| WinoGrande | — | 80.2% |
Math DeepSeek-R1 leads
DeepSeek-R1: 42.0 (#70), Llama 2-70B: 7.5 (#320)
| Benchmark | DeepSeek-R1 | Llama 2-70B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 66.4% | 0% |
| LMArena Math | 1401 | 1091 |
| MATH Level 5 | 96.6% | 3.3% |
| Omni-MATH | 42.4% | — |
| LiveBench Math | 80.7% | — |
| GSM8K | — | 69.6% |
Knowledge DeepSeek-R1 leads
DeepSeek-R1: 44.3 (#81), Llama 2-70B: 7.4 (#305)
| Benchmark | DeepSeek-R1 | Llama 2-70B |
|---|---|---|
| GPQA Diamond | 76.3% | 26.3% |
| LMArena Expert | 1394 | 1039 |
| MMLU-Pro | 79.3% | — |
| Confabulations | 12.7% | — |
| Vectara Hallucination Rate | 11.3% | — |
| GPQA (HELM) | 66.6% | — |
| ARC (AI2) Challenge | — | 78.3% |
| BoolQ | — | 88.6% |
| MMLU | — | 69.9% |
| OpenBookQA | — | 60.2% |
| TriviaQA | — | 87.6% |
Multilingual DeepSeek-R1 leads
DeepSeek-R1: 52.6 (#80), Llama 2-70B: 27.9 (#269)
| Benchmark | DeepSeek-R1 | Llama 2-70B |
|---|---|---|
| LMArena Non-English | 1412 | 1045 |
| LMArena Chinese | 1443 | 994 |
| LMArena French | 1415 | 1089 |
| LMArena German | 1405 | 1041 |
| LMArena Japanese | 1394 | 927 |
| LMArena Korean | 1359 | 964 |
| LMArena Russian | 1423 | 1083 |
| LMArena Spanish | 1410 | 1142 |
Instruction Following DeepSeek-R1 leads
DeepSeek-R1: 72.2 (#138), Llama 2-70B: 55.1 (#273)
| Benchmark | DeepSeek-R1 | Llama 2-70B |
|---|---|---|
| LMArena Instruction Following | 1382 | 1071 |
| LiveBench Instruction Following | 80.5% | — |
| IFEval | 78.4% | — |
Long Context DeepSeek-R1 leads
DeepSeek-R1: 45.4 (#36), Llama 2-70B: 32.2 (#265)
| Benchmark | DeepSeek-R1 | Llama 2-70B |
|---|---|---|
| LMArena Longer Query | 1391 | 1062 |
| Fiction.LiveBench | 75% | — |
Writing & Preference DeepSeek-R1 leads
DeepSeek-R1: 61.5 (#84), Llama 2-70B: 32.5 (#274)
| Benchmark | DeepSeek-R1 | Llama 2-70B |
|---|---|---|
| LMArena Text | 1428 | 1115 |
| LMArena Creative Writing | 1406 | 1075 |
| LMArena Multi-Turn | 1404 | 1088 |
| Short-Story Creative Writing | 83% | — |
| EQ-Bench Creative Writing | 1500 | — |
| WildBench | 82.8% | — |
| LiveBench Language | 48.5% | — |
Frequently asked questions
Is DeepSeek-R1 better than Llama 2-70B?
DeepSeek-R1 is the stronger model overall, scoring 42.0 to 24.1 on the Noometry Index.
Is DeepSeek-R1 or Llama 2-70B better for coding?
DeepSeek-R1 scores higher on coding benchmarks: 45.9 versus 30.8 in the Noometry coding category.
How many benchmarks do DeepSeek-R1 and Llama 2-70B share?
22 benchmarks have published results for both models. DeepSeek-R1 has 52 scored results on Noometry and Llama 2-70B has 35.