Model comparison
DeepSeek-V2 (MoE-236B, May 2024) vs Gemini 3 Pro
Gemini 3 Pro has enough public results to be ranked (#28); DeepSeek-V2 (MoE-236B, May 2024) does not yet, so treat this comparison as directional.
Last verified . 1 shared benchmarks.
DeepSeek
GoogleSummary
- They share 1 benchmark with published results for both. DeepSeek-V2 (MoE-236B, May 2024) scores higher in 0 categories and Gemini 3 Pro in 1 category; one gap is clear of the uncertainty.
- The widest gap is in coding, where Gemini 3 Pro leads 51.6 to 40.4.
- DeepSeek-V2 (MoE-236B, May 2024) has downloadable open weights; the other is API-only.
Side by side
| DeepSeek-V2 (MoE-236B, May 2024) | Gemini 3 Pro | |
|---|---|---|
| Provider | DeepSeek | |
| Noometry Index | 40.3 | 54.8 |
| Released | 2024-05-07 | 2025-11-18 |
| Weights | Open | Proprietary |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 10 | 67 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Gemini 3 Pro leads
DeepSeek-V2 (MoE-236B, May 2024): 40.4 (#139), Gemini 3 Pro: 51.6 (#39)
| Benchmark | DeepSeek-V2 (MoE-236B, May 2024) | Gemini 3 Pro |
|---|---|---|
| SWE-bench Verified | — | 72.9% |
| SWE-bench Verified (bash only) | — | 74.2% |
| LMArena WebDev | — | 1440 |
| SWE-bench Multilingual | — | 68.7% |
| GSO | — | 18.6% |
| WeirdML | — | 69.9% |
| BigCodeBench Instruct | 48.9% | — |
| LMArena Coding | — | 1481 |
| BigCodeBench Complete | 59.4% | — |
| ALE-Bench | — | 1,177 |
| AlgoTune | — | 1.83 |
Agentic & Tool Use Not comparable
DeepSeek-V2 (MoE-236B, May 2024): —, Gemini 3 Pro: 40.6 (#23)
| Benchmark | DeepSeek-V2 (MoE-236B, May 2024) | Gemini 3 Pro |
|---|---|---|
| Terminal-Bench | — | 69.4% |
| Berkeley Function Calling Leaderboard | — | 72.5% |
| GDPval | — | 40.3% |
| Remote Labor Index | — | 1.3% |
| τ²-bench Airline | — | 80.5% |
| τ²-bench Banking | — | 18% |
| τ²-bench Retail | — | 75.9% |
| τ²-bench Telecom | — | 91% |
| DeepResearch Bench | — | 46.3% |
| BALROG | — | 58.1% |
| LMArena Search | — | 1207 |
| METR Time Horizons | — | 71% |
| Vending-Bench 2 | — | 5,478 |
Reasoning Not comparable
DeepSeek-V2 (MoE-236B, May 2024): —, Gemini 3 Pro: 52.5 (#31)
| Benchmark | DeepSeek-V2 (MoE-236B, May 2024) | Gemini 3 Pro |
|---|---|---|
| Epoch Capabilities Index | 124.77 | 152.92 |
| ARC-AGI-2 | — | 31.1% |
| SimpleBench | — | 76.4% |
| Kagi LLM Benchmark | — | 80.1% |
| NYT Connections (extended) | — | 94.4% |
| ARC-AGI-1 | — | 75% |
| CritPt | — | 6.9% |
| Chess Puzzles | — | 31% |
| EnigmaEval | — | 18.2% |
| LMArena Hard Prompts | — | 1480 |
| BIG-Bench Hard | 78.8% | — |
| ForecastBench | — | 61.2 |
| HellaSwag | 87.1% | — |
| PIQA | 83.9% | — |
| WinoGrande | 86.3% | — |
Math Not comparable
DeepSeek-V2 (MoE-236B, May 2024): —, Gemini 3 Pro: 49.9 (#59)
| Benchmark | DeepSeek-V2 (MoE-236B, May 2024) | Gemini 3 Pro |
|---|---|---|
| MathArena Final-Answer Competitions | — | 67% |
| OTIS Mock AIME 2024-2025 | — | 91.4% |
| ProofBench | — | 20% |
| Omni-MATH | — | 55.5% |
| LMArena Math | — | 1476 |
| FrontierMath (Feb 2025 set) | — | 37.6% |
| FrontierMath Tier 4 (v1) | — | 18.8% |
Knowledge Not comparable
DeepSeek-V2 (MoE-236B, May 2024): —, Gemini 3 Pro: 64.4 (#16)
| Benchmark | DeepSeek-V2 (MoE-236B, May 2024) | Gemini 3 Pro |
|---|---|---|
| GPQA Diamond | — | 92.6% |
| Humanity's Last Exam | — | 37.5% |
| MMLU-Pro | — | 90.3% |
| Vectara Hallucination Rate | — | 13.6% |
| GPQA (HELM) | — | 80.3% |
| LMArena Expert | — | 1475 |
| ARC (AI2) Challenge | 92.2% | — |
| MMLU | 78.4% | — |
| TriviaQA | 80% | — |
Multimodal Not comparable
DeepSeek-V2 (MoE-236B, May 2024): —, Gemini 3 Pro: 57.6 (#2)
| Benchmark | DeepSeek-V2 (MoE-236B, May 2024) | Gemini 3 Pro |
|---|---|---|
| LMArena Vision | — | 1305 |
| GeoBench | — | 84% |
| VPCT | — | 91% |
| LMArena Document | — | 1434 |
Multilingual Not comparable
DeepSeek-V2 (MoE-236B, May 2024): —, Gemini 3 Pro: 56.9 (#16)
| Benchmark | DeepSeek-V2 (MoE-236B, May 2024) | Gemini 3 Pro |
|---|---|---|
| LMArena Non-English | — | 1474 |
| LMArena Chinese | — | 1523 |
| LMArena French | — | 1492 |
| LMArena German | — | 1515 |
| LMArena Japanese | — | 1510 |
| LMArena Korean | — | 1448 |
| LMArena Russian | — | 1493 |
| LMArena Spanish | — | 1470 |
Instruction Following Not comparable
DeepSeek-V2 (MoE-236B, May 2024): —, Gemini 3 Pro: 76.3 (#45)
| Benchmark | DeepSeek-V2 (MoE-236B, May 2024) | Gemini 3 Pro |
|---|---|---|
| IFEval | — | 87.7% |
| LMArena Instruction Following | — | 1458 |
Long Context Not comparable
DeepSeek-V2 (MoE-236B, May 2024): —, Gemini 3 Pro: 44.0 (#79)
| Benchmark | DeepSeek-V2 (MoE-236B, May 2024) | Gemini 3 Pro |
|---|---|---|
| CL-bench | — | 15.8% |
| LMArena Longer Query | — | 1471 |
Writing & Preference Not comparable
DeepSeek-V2 (MoE-236B, May 2024): —, Gemini 3 Pro: 66.4 (#35)
| Benchmark | DeepSeek-V2 (MoE-236B, May 2024) | Gemini 3 Pro |
|---|---|---|
| LMArena Text | — | 1479 |
| LMArena Creative Writing | — | 1482 |
| EQ-Bench Creative Writing | — | 1525 |
| WildBench | — | 85.9% |
| LMArena Multi-Turn | — | 1484 |
Frequently asked questions
Is DeepSeek-V2 (MoE-236B, May 2024) better than Gemini 3 Pro?
Gemini 3 Pro has enough public results to be ranked (#28); DeepSeek-V2 (MoE-236B, May 2024) does not yet, so treat this comparison as directional.
Is DeepSeek-V2 (MoE-236B, May 2024) or Gemini 3 Pro better for coding?
Gemini 3 Pro scores higher on coding benchmarks: 51.6 versus 40.4 in the Noometry coding category.
How many benchmarks do DeepSeek-V2 (MoE-236B, May 2024) and Gemini 3 Pro share?
1 benchmark has published results for both models. DeepSeek-V2 (MoE-236B, May 2024) has 10 scored results on Noometry and Gemini 3 Pro has 67.