Model comparison
DeepSeek-V2 (MoE-236B, May 2024) vs Llama 13b
Llama 13b has enough public results to be ranked (#348); 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 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 21.4.
Side by side
| DeepSeek-V2 (MoE-236B, May 2024) | Llama 13b | |
|---|---|---|
| Provider | DeepSeek | Meta |
| Noometry Index | 40.3 | 24.4 |
| Released | 2024-05-07 | 2023-02-24 |
| Weights | Open | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 10 | 21 |
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 13b: 21.4 (#337)
| Benchmark | DeepSeek-V2 (MoE-236B, May 2024) | Llama 13b |
|---|---|---|
| BigCodeBench Instruct | 48.9% | — |
| LMArena Coding | — | 683 |
| BigCodeBench Complete | 59.4% | — |
Reasoning Not comparable
DeepSeek-V2 (MoE-236B, May 2024): —, Llama 13b: 14.0 (#329)
| Benchmark | DeepSeek-V2 (MoE-236B, May 2024) | Llama 13b |
|---|---|---|
| BIG-Bench Hard | 78.8% | 37.9% |
| Epoch Capabilities Index | 124.77 | 100.58 |
| HellaSwag | 87.1% | 79.2% |
| PIQA | 83.9% | 80.1% |
| WinoGrande | 86.3% | 73% |
| LMArena Hard Prompts | — | 728 |
| LAMBADA | — | 75.2% |
Math Not comparable
DeepSeek-V2 (MoE-236B, May 2024): —, Llama 13b: 26.7 (#256)
| Benchmark | DeepSeek-V2 (MoE-236B, May 2024) | Llama 13b |
|---|---|---|
| LMArena Math | — | 838 |
| GSM8K | — | 20.6% |
Knowledge Not comparable
DeepSeek-V2 (MoE-236B, May 2024): —, Llama 13b: —
| Benchmark | DeepSeek-V2 (MoE-236B, May 2024) | Llama 13b |
|---|---|---|
| ARC (AI2) Challenge | 92.2% | 52.7% |
| MMLU | 78.4% | 47.7% |
| TriviaQA | 80% | 77.9% |
| BoolQ | — | 78.7% |
| OpenBookQA | — | 56.4% |
Multimodal Not comparable
DeepSeek-V2 (MoE-236B, May 2024): —, Llama 13b: —
| Benchmark | DeepSeek-V2 (MoE-236B, May 2024) | Llama 13b |
|---|---|---|
| ScienceQA | — | 43.3% |
Multilingual Not comparable
DeepSeek-V2 (MoE-236B, May 2024): —, Llama 13b: 16.6 (#297)
| Benchmark | DeepSeek-V2 (MoE-236B, May 2024) | Llama 13b |
|---|---|---|
| LMArena Non-English | — | 819 |
Instruction Following Not comparable
DeepSeek-V2 (MoE-236B, May 2024): —, Llama 13b: 36.7 (#305)
| Benchmark | DeepSeek-V2 (MoE-236B, May 2024) | Llama 13b |
|---|---|---|
| LMArena Instruction Following | — | 781 |
Writing & Preference Not comparable
DeepSeek-V2 (MoE-236B, May 2024): —, Llama 13b: 13.8 (#312)
| Benchmark | DeepSeek-V2 (MoE-236B, May 2024) | Llama 13b |
|---|---|---|
| LMArena Text | — | 834 |
| LMArena Creative Writing | — | 794 |
| LMArena Multi-Turn | — | 753 |
Frequently asked questions
Is DeepSeek-V2 (MoE-236B, May 2024) better than Llama 13b?
Llama 13b has enough public results to be ranked (#348); DeepSeek-V2 (MoE-236B, May 2024) does not yet, so treat this comparison as directional.
Is DeepSeek-V2 (MoE-236B, May 2024) or Llama 13b better for coding?
DeepSeek-V2 (MoE-236B, May 2024) scores higher on coding benchmarks: 40.4 versus 21.4 in the Noometry coding category.
How many benchmarks do DeepSeek-V2 (MoE-236B, May 2024) and Llama 13b share?
8 benchmarks have published results for both models. DeepSeek-V2 (MoE-236B, May 2024) has 10 scored results on Noometry and Llama 13b has 21.