Model comparison
DeepSeek-R1 vs Llama 3.1-405B
DeepSeek-R1 is the stronger model overall, scoring 42.3 to 30.7 on the Noometry Index.
Last verified . 32 shared benchmarks.
DeepSeek
MetaSummary
- They share 32 benchmarks with published results for both. DeepSeek-R1 scores higher in 9 categories and Llama 3.1-405B in 0 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where DeepSeek-R1 leads 43.8 to 18.4.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 66.4% for DeepSeek-R1 and 9.7% for Llama 3.1-405B.
- Llama 3.1-405B has downloadable open weights; the other is API-only.
Side by side
| DeepSeek-R1 | Llama 3.1-405B | |
|---|---|---|
| Provider | DeepSeek | Meta |
| Noometry Index | 42.3 | 30.7 |
| Released | 2025-01-20 | 2024-07-23 |
| Weights | Proprietary | Open |
| Context window | 164K | — |
| Max output | 64K | — |
| Input $ / M tokens | $0.50 | — |
| Output $ / M tokens | $2.15 | — |
| Results tracked | 52 | 42 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek-R1 leads
DeepSeek-R1: 46.3 (#68), Llama 3.1-405B: 33.1 (#262)
| Benchmark | DeepSeek-R1 | Llama 3.1-405B |
|---|---|---|
| WeirdML | 41.6% | 21.4% |
| LMArena Coding | 1427 | 1291 |
| Aider Polyglot | 71.4% | — |
| SciCode | 35.7% | — |
| LiveBench Coding | 66.7% | — |
| ALE-Bench | 804.12 | — |
| AlgoTune | 1.7 | — |
Agentic & Tool Use DeepSeek-R1 leads
DeepSeek-R1: 30.7 (#75), Llama 3.1-405B: 21.0 (#140)
| Benchmark | DeepSeek-R1 | Llama 3.1-405B |
|---|---|---|
| TheAgentCompany | — | 7.4% |
| Cybench | — | 7.5% |
| DeepResearch Bench | 35.1% | — |
| BALROG | 34.9% | — |
| METR Time Horizons | 53.8% | — |
Reasoning DeepSeek-R1 leads
DeepSeek-R1: 18.6 (#278), Llama 3.1-405B: 16.8 (#300)
| Benchmark | DeepSeek-R1 | Llama 3.1-405B |
|---|---|---|
| SimpleBench | 40.8% | 23% |
| Kagi LLM Benchmark | 69.4% | 45% |
| LMArena Hard Prompts | 1416 | 1269 |
| Epoch Capabilities Index | 141.29 | 128.75 |
| ForecastBench | 60 | 59.9 |
| ARC-AGI-2 | 1.3% | — |
| ARC-AGI-1 | 21.2% | — |
| CritPt | 1.1% | — |
| LiveBench Reasoning | 83.2% | — |
| DTBench | — | 61.4% |
| LiveBench Data Analysis | 69.8% | — |
| BIG-Bench Hard | — | 82.9% |
| HellaSwag | — | 89.2% |
| LiveBench | 71.6% | — |
| PIQA | — | 85.9% |
| WinoGrande | — | 89.2% |
Math DeepSeek-R1 leads
DeepSeek-R1: 43.8 (#79), Llama 3.1-405B: 18.4 (#290)
| Benchmark | DeepSeek-R1 | Llama 3.1-405B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 66.4% | 9.7% |
| Omni-MATH | 42.4% | 24.9% |
| LMArena Math | 1400 | 1281 |
| MATH Level 5 | 96.6% | 49.8% |
| LiveBench Math | 80.7% | — |
Knowledge DeepSeek-R1 leads
DeepSeek-R1: 44.5 (#87), Llama 3.1-405B: 30.4 (#227)
| Benchmark | DeepSeek-R1 | Llama 3.1-405B |
|---|---|---|
| GPQA Diamond | 76.3% | 50.9% |
| MMLU-Pro | 79.3% | 72.3% |
| Confabulations | 12.7% | 17.6% |
| GPQA (HELM) | 66.6% | 52.2% |
| LMArena Expert | 1394 | 1243 |
| Vectara Hallucination Rate | 11.3% | — |
| ARC (AI2) Challenge | — | 95.3% |
| MMLU | — | 84.5% |
| TriviaQA | — | 82.7% |
Multilingual DeepSeek-R1 leads
DeepSeek-R1: 52.4 (#85), Llama 3.1-405B: 40.7 (#214)
| Benchmark | DeepSeek-R1 | Llama 3.1-405B |
|---|---|---|
| LMArena Non-English | 1412 | 1248 |
| LMArena Chinese | 1442 | 1242 |
| LMArena French | 1417 | 1279 |
| LMArena German | 1404 | 1252 |
| LMArena Japanese | 1391 | 1208 |
| LMArena Korean | 1360 | 1184 |
| LMArena Russian | 1423 | 1265 |
| LMArena Spanish | 1411 | 1260 |
Instruction Following DeepSeek-R1 leads
DeepSeek-R1: 72.0 (#143), Llama 3.1-405B: 65.9 (#214)
| Benchmark | DeepSeek-R1 | Llama 3.1-405B |
|---|---|---|
| IFEval | 78.4% | 81.1% |
| LMArena Instruction Following | 1382 | 1259 |
| LiveBench Instruction Following | 80.5% | — |
Long Context DeepSeek-R1 leads
DeepSeek-R1: 45.4 (#36), Llama 3.1-405B: 38.4 (#197)
| Benchmark | DeepSeek-R1 | Llama 3.1-405B |
|---|---|---|
| LMArena Longer Query | 1391 | 1266 |
| Fiction.LiveBench | 75% | — |
Writing & Preference DeepSeek-R1 leads
DeepSeek-R1: 61.4 (#88), Llama 3.1-405B: 38.9 (#251)
| Benchmark | DeepSeek-R1 | Llama 3.1-405B |
|---|---|---|
| LMArena Text | 1428 | 1284 |
| LMArena Creative Writing | 1405 | 1262 |
| EQ-Bench Creative Writing | 1500 | 870 |
| WildBench | 82.8% | 78.3% |
| LMArena Multi-Turn | 1405 | 1297 |
| Short-Story Creative Writing | 83% | — |
| LiveBench Language | 48.5% | — |
Frequently asked questions
Is DeepSeek-R1 better than Llama 3.1-405B?
DeepSeek-R1 is the stronger model overall, scoring 42.3 to 30.7 on the Noometry Index.
Is DeepSeek-R1 or Llama 3.1-405B better for coding?
DeepSeek-R1 scores higher on coding benchmarks: 46.3 versus 33.1 in the Noometry coding category.
How many benchmarks do DeepSeek-R1 and Llama 3.1-405B share?
32 benchmarks have published results for both models. DeepSeek-R1 has 52 scored results on Noometry and Llama 3.1-405B has 42.