Model comparison
DeepSeek-V2.5 (Sep 2024) vs Llama 3.1-405B
DeepSeek-V2.5 (Sep 2024) is the stronger model overall, scoring 37.0 to 30.4 on the Noometry Index.
Last verified . 17 shared benchmarks.
DeepSeek
MetaSummary
- They share 17 benchmarks with published results for both. DeepSeek-V2.5 (Sep 2024) scores higher in 7 categories and Llama 3.1-405B in 1 category; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where DeepSeek-V2.5 (Sep 2024) leads 33.3 to 17.3.
Side by side
| DeepSeek-V2.5 (Sep 2024) | Llama 3.1-405B | |
|---|---|---|
| Provider | DeepSeek | Meta |
| Noometry Index | 37.0 | 30.4 |
| Released | 2024-09-06 | 2024-07-23 |
| Weights | Open | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 22 | 42 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Llama 3.1-405B leads
DeepSeek-V2.5 (Sep 2024): 31.3 (#274), Llama 3.1-405B: 32.7 (#257)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Llama 3.1-405B |
|---|---|---|
| LMArena Coding | 1310 | 1291 |
| Aider Polyglot | 17.8% | — |
| WeirdML | — | 21.4% |
| BigCodeBench Instruct | 48.6% | — |
| BigCodeBench Complete | 53.2% | — |
| HumanEval+ | 83.5% | — |
| MBPP+ | 74.1% | — |
Agentic & Tool Use Not comparable
DeepSeek-V2.5 (Sep 2024): —, Llama 3.1-405B: 20.9 (#131)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Llama 3.1-405B |
|---|---|---|
| TheAgentCompany | — | 7.4% |
| Cybench | — | 7.5% |
Reasoning DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 24.5 (#145), Llama 3.1-405B: 16.4 (#293)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Llama 3.1-405B |
|---|---|---|
| LMArena Hard Prompts | 1289 | 1270 |
| SimpleBench | — | 23% |
| Kagi LLM Benchmark | — | 45% |
| DTBench | — | 61.4% |
| BIG-Bench Hard | — | 82.9% |
| Epoch Capabilities Index | — | 128.76 |
| ForecastBench | — | 59.9 |
| HellaSwag | — | 89.2% |
| PIQA | — | 85.9% |
| WinoGrande | — | 89.2% |
Math DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 33.3 (#177), Llama 3.1-405B: 17.3 (#283)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Llama 3.1-405B |
|---|---|---|
| LMArena Math | 1289 | 1281 |
| OTIS Mock AIME 2024-2025 | — | 9.7% |
| Omni-MATH | — | 24.9% |
| MATH Level 5 | — | 49.8% |
Knowledge DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 34.2 (#190), Llama 3.1-405B: 30.4 (#221)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Llama 3.1-405B |
|---|---|---|
| LMArena Expert | 1266 | 1243 |
| GPQA Diamond | — | 50.9% |
| MMLU-Pro | — | 72.3% |
| Confabulations | — | 17.6% |
| GPQA (HELM) | — | 52.2% |
| ARC (AI2) Challenge | — | 95.3% |
| MMLU | — | 84.5% |
| TriviaQA | — | 82.7% |
Multilingual DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 42.7 (#188), Llama 3.1-405B: 40.9 (#209)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Llama 3.1-405B |
|---|---|---|
| LMArena Non-English | 1273 | 1248 |
| LMArena Chinese | 1317 | 1242 |
| LMArena French | 1288 | 1278 |
| LMArena German | 1258 | 1252 |
| LMArena Japanese | 1228 | 1208 |
| LMArena Korean | 1209 | 1184 |
| LMArena Russian | 1289 | 1265 |
| LMArena Spanish | 1248 | 1259 |
Instruction Following DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 67.7 (#189), Llama 3.1-405B: 66.1 (#209)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Llama 3.1-405B |
|---|---|---|
| LMArena Instruction Following | 1280 | 1259 |
| IFEval | — | 81.1% |
Long Context DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 39.5 (#169), Llama 3.1-405B: 38.4 (#192)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Llama 3.1-405B |
|---|---|---|
| LMArena Longer Query | 1301 | 1266 |
Writing & Preference DeepSeek-V2.5 (Sep 2024) leads
DeepSeek-V2.5 (Sep 2024): 50.0 (#182), Llama 3.1-405B: 39.1 (#246)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Llama 3.1-405B |
|---|---|---|
| LMArena Text | 1294 | 1284 |
| LMArena Creative Writing | 1285 | 1262 |
| LMArena Multi-Turn | 1297 | 1297 |
| EQ-Bench Creative Writing | — | 870 |
| WildBench | — | 78.3% |
Frequently asked questions
Is DeepSeek-V2.5 (Sep 2024) better than Llama 3.1-405B?
DeepSeek-V2.5 (Sep 2024) is the stronger model overall, scoring 37.0 to 30.4 on the Noometry Index.
Is DeepSeek-V2.5 (Sep 2024) or Llama 3.1-405B better for coding?
Llama 3.1-405B scores higher on coding benchmarks: 32.7 versus 31.3 in the Noometry coding category.
How many benchmarks do DeepSeek-V2.5 (Sep 2024) and Llama 3.1-405B share?
17 benchmarks have published results for both models. DeepSeek-V2.5 (Sep 2024) has 22 scored results on Noometry and Llama 3.1-405B has 42.