Model comparison
DeepSeek-V2 (MoE-236B, May 2024) vs Gemma 2B
Gemma 2B has enough public results to be ranked (#307); DeepSeek-V2 (MoE-236B, May 2024) does not yet, so treat this comparison as directional.
Last verified . 8 shared benchmarks.
DeepSeek
GoogleSummary
- They share 8 benchmarks with published results for both. DeepSeek-V2 (MoE-236B, May 2024) scores higher in 1 category and Gemma 2B 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 29.4.
Side by side
| DeepSeek-V2 (MoE-236B, May 2024) | Gemma 2B | |
|---|---|---|
| Provider | DeepSeek | |
| Noometry Index | 40.3 | 29.6 |
| Released | 2024-05-07 | 2024-02-21 |
| Weights | Open | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 10 | 23 |
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), Gemma 2B: 29.4 (#305)
| Benchmark | DeepSeek-V2 (MoE-236B, May 2024) | Gemma 2B |
|---|---|---|
| BigCodeBench Instruct | 48.9% | — |
| LMArena Coding | — | 1010 |
| BigCodeBench Complete | 59.4% | — |
| HumanEval+ | — | 20.7% |
| MBPP+ | — | 34.1% |
Reasoning Not comparable
DeepSeek-V2 (MoE-236B, May 2024): —, Gemma 2B: 18.8 (#275)
| Benchmark | DeepSeek-V2 (MoE-236B, May 2024) | Gemma 2B |
|---|---|---|
| BIG-Bench Hard | 78.8% | 35.2% |
| Epoch Capabilities Index | 124.77 | 94.2 |
| HellaSwag | 87.1% | 71.4% |
| PIQA | 83.9% | 77.3% |
| WinoGrande | 86.3% | 65.4% |
| LMArena Hard Prompts | — | 989 |
Math Not comparable
DeepSeek-V2 (MoE-236B, May 2024): —, Gemma 2B: 30.0 (#239)
| Benchmark | DeepSeek-V2 (MoE-236B, May 2024) | Gemma 2B |
|---|---|---|
| LMArena Math | — | 1009 |
| GSM8K | — | 17.7% |
Knowledge Not comparable
DeepSeek-V2 (MoE-236B, May 2024): —, Gemma 2B: —
| Benchmark | DeepSeek-V2 (MoE-236B, May 2024) | Gemma 2B |
|---|---|---|
| ARC (AI2) Challenge | 92.2% | 42.1% |
| MMLU | 78.4% | 42.3% |
| TriviaQA | 80% | 53.2% |
| BoolQ | — | 69.4% |
Multilingual Not comparable
DeepSeek-V2 (MoE-236B, May 2024): —, Gemma 2B: 23.0 (#294)
| Benchmark | DeepSeek-V2 (MoE-236B, May 2024) | Gemma 2B |
|---|---|---|
| LMArena Non-English | — | 958 |
| LMArena Chinese | — | 986 |
| LMArena Russian | — | 937 |
Instruction Following Not comparable
DeepSeek-V2 (MoE-236B, May 2024): —, Gemma 2B: 48.5 (#302)
| Benchmark | DeepSeek-V2 (MoE-236B, May 2024) | Gemma 2B |
|---|---|---|
| LMArena Instruction Following | — | 970 |
Long Context Not comparable
DeepSeek-V2 (MoE-236B, May 2024): —, Gemma 2B: 29.9 (#291)
| Benchmark | DeepSeek-V2 (MoE-236B, May 2024) | Gemma 2B |
|---|---|---|
| LMArena Longer Query | — | 981 |
Writing & Preference Not comparable
DeepSeek-V2 (MoE-236B, May 2024): —, Gemma 2B: 24.0 (#308)
| Benchmark | DeepSeek-V2 (MoE-236B, May 2024) | Gemma 2B |
|---|---|---|
| LMArena Text | — | 1002 |
| LMArena Creative Writing | — | 987 |
| LMArena Multi-Turn | — | 945 |
Frequently asked questions
Is DeepSeek-V2 (MoE-236B, May 2024) better than Gemma 2B?
Gemma 2B has enough public results to be ranked (#307); DeepSeek-V2 (MoE-236B, May 2024) does not yet, so treat this comparison as directional.
Is DeepSeek-V2 (MoE-236B, May 2024) or Gemma 2B better for coding?
DeepSeek-V2 (MoE-236B, May 2024) scores higher on coding benchmarks: 40.4 versus 29.4 in the Noometry coding category.
How many benchmarks do DeepSeek-V2 (MoE-236B, May 2024) and Gemma 2B share?
8 benchmarks have published results for both models. DeepSeek-V2 (MoE-236B, May 2024) has 10 scored results on Noometry and Gemma 2B has 23.