Model comparison
DeepSeek-V2 (MoE-236B, May 2024) vs Llama 2-13B
Llama 2-13B has enough public results to be ranked (#309); DeepSeek-V2 (MoE-236B, May 2024) does not yet, so treat this comparison as directional.
Last verified . 8 shared benchmarks.
DeepSeek
MetaSummary
- They share 8 benchmarks with published results for both. DeepSeek-V2 (MoE-236B, May 2024) scores higher in 1 category and Llama 2-13B in 0 categories; one gap is clear of the uncertainty.
- The widest gap is in coding, where DeepSeek-V2 (MoE-236B, May 2024) leads 40.4 to 30.9.
Side by side
| DeepSeek-V2 (MoE-236B, May 2024) | Llama 2-13B | |
|---|---|---|
| Provider | DeepSeek | Meta |
| Noometry Index | 40.3 | 29.6 |
| Released | 2024-05-07 | 2023-07-18 |
| Weights | Open | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 10 | 32 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek-V2 (MoE-236B, May 2024) leads
DeepSeek-V2 (MoE-236B, May 2024): 40.4 (#139), Llama 2-13B: 30.9 (#291)
| Benchmark | DeepSeek-V2 (MoE-236B, May 2024) | Llama 2-13B |
|---|---|---|
| BigCodeBench Instruct | 48.9% | — |
| LMArena Coding | — | 1062 |
| BigCodeBench Complete | 59.4% | — |
Reasoning Not comparable
DeepSeek-V2 (MoE-236B, May 2024): —, Llama 2-13B: 12.8 (#337)
| Benchmark | DeepSeek-V2 (MoE-236B, May 2024) | Llama 2-13B |
|---|---|---|
| BIG-Bench Hard | 78.8% | 58.2% |
| Epoch Capabilities Index | 124.77 | 106.17 |
| HellaSwag | 87.1% | 80.7% |
| PIQA | 83.9% | 80.8% |
| WinoGrande | 86.3% | 72.8% |
| Chess Puzzles | — | 0% |
| LMArena Hard Prompts | — | 1051 |
| DTBench | — | 42.2% |
| LAMBADA | — | 76.5% |
Math Not comparable
DeepSeek-V2 (MoE-236B, May 2024): —, Llama 2-13B: 31.1 (#229)
| Benchmark | DeepSeek-V2 (MoE-236B, May 2024) | Llama 2-13B |
|---|---|---|
| LMArena Math | — | 1065 |
| GSM8K | — | 36.9% |
Knowledge Not comparable
DeepSeek-V2 (MoE-236B, May 2024): —, Llama 2-13B: 28.1 (#249)
| Benchmark | DeepSeek-V2 (MoE-236B, May 2024) | Llama 2-13B |
|---|---|---|
| ARC (AI2) Challenge | 92.2% | 60.3% |
| MMLU | 78.4% | 55.6% |
| TriviaQA | 80% | 79.6% |
| LMArena Expert | — | 1030 |
| BoolQ | — | 82.4% |
| OpenBookQA | — | 57% |
Multimodal Not comparable
DeepSeek-V2 (MoE-236B, May 2024): —, Llama 2-13B: —
| Benchmark | DeepSeek-V2 (MoE-236B, May 2024) | Llama 2-13B |
|---|---|---|
| ScienceQA | — | 55.8% |
Multilingual Not comparable
DeepSeek-V2 (MoE-236B, May 2024): —, Llama 2-13B: 26.5 (#279)
| Benchmark | DeepSeek-V2 (MoE-236B, May 2024) | Llama 2-13B |
|---|---|---|
| LMArena Non-English | — | 1024 |
| LMArena Chinese | — | 1001 |
| LMArena French | — | 1044 |
| LMArena German | — | 1009 |
| LMArena Japanese | — | 894 |
| LMArena Korean | — | 953 |
| LMArena Russian | — | 1055 |
| LMArena Spanish | — | 1087 |
Instruction Following Not comparable
DeepSeek-V2 (MoE-236B, May 2024): —, Llama 2-13B: 53.3 (#287)
| Benchmark | DeepSeek-V2 (MoE-236B, May 2024) | Llama 2-13B |
|---|---|---|
| LMArena Instruction Following | — | 1045 |
Long Context Not comparable
DeepSeek-V2 (MoE-236B, May 2024): —, Llama 2-13B: 32.3 (#269)
| Benchmark | DeepSeek-V2 (MoE-236B, May 2024) | Llama 2-13B |
|---|---|---|
| LMArena Longer Query | — | 1064 |
Writing & Preference Not comparable
DeepSeek-V2 (MoE-236B, May 2024): —, Llama 2-13B: 29.8 (#289)
| Benchmark | DeepSeek-V2 (MoE-236B, May 2024) | Llama 2-13B |
|---|---|---|
| LMArena Text | — | 1084 |
| LMArena Creative Writing | — | 1047 |
| LMArena Multi-Turn | — | 1050 |
Frequently asked questions
Is DeepSeek-V2 (MoE-236B, May 2024) better than Llama 2-13B?
Llama 2-13B has enough public results to be ranked (#309); DeepSeek-V2 (MoE-236B, May 2024) does not yet, so treat this comparison as directional.
Is DeepSeek-V2 (MoE-236B, May 2024) or Llama 2-13B better for coding?
DeepSeek-V2 (MoE-236B, May 2024) scores higher on coding benchmarks: 40.4 versus 30.9 in the Noometry coding category.
How many benchmarks do DeepSeek-V2 (MoE-236B, May 2024) and Llama 2-13B share?
8 benchmarks have published results for both models. DeepSeek-V2 (MoE-236B, May 2024) has 10 scored results on Noometry and Llama 2-13B has 32.