Model comparison
DeepSeek-V3 vs Gemini 2.0 Flash (Feb 2025)
DeepSeek-V3 is the stronger model overall, scoring 39.5 to 35.1 on the Noometry Index.
Last verified . 46 shared benchmarks.
DeepSeek
GoogleSummary
- They share 46 benchmarks with published results for both. DeepSeek-V3 scores higher in 5 categories and Gemini 2.0 Flash (Feb 2025) in 3 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in coding, where DeepSeek-V3 leads 42.3 to 28.4.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 37.8% for DeepSeek-V3 and 57.8% for Gemini 2.0 Flash (Feb 2025).
- DeepSeek-V3 has downloadable open weights; the other is API-only.
Side by side
| DeepSeek-V3 | Gemini 2.0 Flash (Feb 2025) | |
|---|---|---|
| Provider | DeepSeek | |
| Noometry Index | 39.5 | 35.1 |
| Released | 2024-12-26 | 2024-12-06 |
| Weights | Open | Proprietary |
| Context window | 164K | — |
| Max output | 164K | — |
| Input $ / M tokens | $0.24 | — |
| Output $ / M tokens | $0.90 | — |
| Results tracked | 60 | 54 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek-V3 leads
DeepSeek-V3: 42.3 (#106), Gemini 2.0 Flash (Feb 2025): 28.4 (#315)
| Benchmark | DeepSeek-V3 | Gemini 2.0 Flash (Feb 2025) |
|---|---|---|
| Aider Polyglot | 55.1% | 38.2% |
| WeirdML | 36.1% | 25.8% |
| BigCodeBench Instruct | 50% | 45.9% |
| LiveBench Coding | 70.9% | 63.4% |
| LMArena Coding | 1368 | 1350 |
| BigCodeBench Complete | 62.2% | 59.9% |
| SWE-bench Verified (bash only) | — | 13.5% |
| SciCode | 35.8% | — |
| CadEval | — | 30% |
| HumanEval+ | 86.6% | — |
| MBPP+ | 73% | — |
Agentic & Tool Use Not comparable
DeepSeek-V3: —, Gemini 2.0 Flash (Feb 2025): 28.1 (#92)
| Benchmark | DeepSeek-V3 | Gemini 2.0 Flash (Feb 2025) |
|---|---|---|
| TheAgentCompany | — | 11.4% |
| METR Time Horizons | 49.6% | — |
Reasoning DeepSeek-V3 leads
DeepSeek-V3: 20.5 (#236), Gemini 2.0 Flash (Feb 2025): 15.2 (#318)
| Benchmark | DeepSeek-V3 | Gemini 2.0 Flash (Feb 2025) |
|---|---|---|
| SimpleBench | 27.2% | 31.1% |
| Kagi LLM Benchmark | 52.3% | 37.8% |
| LiveBench Reasoning | 65.8% | 78.2% |
| LMArena Hard Prompts | 1365 | 1346 |
| DTBench | 64.8% | 63.2% |
| LiveBench Data Analysis | 60.9% | 69.4% |
| Epoch Capabilities Index | 135.94 | 135.36 |
| LiveBench | 66.9% | 66.9% |
| ARC-AGI-2 | — | 1.3% |
| CritPt | 0% | — |
| EnigmaEval | — | 1.1% |
| LMCA | 15.5% | — |
| BIG-Bench Hard | 87.5% | — |
| ForecastBench | 59.1 | — |
| HellaSwag | 88.9% | — |
| PIQA | 84.7% | — |
| WinoGrande | 85.2% | — |
Math Gemini 2.0 Flash (Feb 2025) leads
DeepSeek-V3: 32.1 (#219), Gemini 2.0 Flash (Feb 2025): 37.9 (#146)
| Benchmark | DeepSeek-V3 | Gemini 2.0 Flash (Feb 2025) |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 37.8% | 57.8% |
| Omni-MATH | 40.3% | 45.9% |
| LiveBench Math | 73.5% | 75.8% |
| LMArena Math | 1373 | 1352 |
| MATH Level 5 | 75.5% | 82.2% |
| FrontierMath (Feb 2025 set) | 1.7% | 1.7% |
Knowledge DeepSeek-V3 leads
DeepSeek-V3: 37.5 (#155), Gemini 2.0 Flash (Feb 2025): 32.0 (#213)
| Benchmark | DeepSeek-V3 | Gemini 2.0 Flash (Feb 2025) |
|---|---|---|
| GPQA Diamond | 67.6% | 64.1% |
| MMLU-Pro | 72.3% | 73.7% |
| Confabulations | 26.1% | 12.4% |
| GPQA (HELM) | 53.8% | 55.6% |
| LMArena Expert | 1351 | 1339 |
| MMLU | 87.2% | 79.7% |
| Humanity's Last Exam | — | 6.6% |
| Vectara Hallucination Rate | 6.1% | — |
| ARC (AI2) Challenge | 95.3% | — |
| TriviaQA | 82.9% | — |
Multimodal Not comparable
DeepSeek-V3: —, Gemini 2.0 Flash (Feb 2025): 36.5 (#79)
| Benchmark | DeepSeek-V3 | Gemini 2.0 Flash (Feb 2025) |
|---|---|---|
| LMArena Vision | — | 1158 |
| GeoBench | — | 77% |
Multilingual DeepSeek-V3 leads
DeepSeek-V3: 48.5 (#143), Gemini 2.0 Flash (Feb 2025): 47.4 (#149)
| Benchmark | DeepSeek-V3 | Gemini 2.0 Flash (Feb 2025) |
|---|---|---|
| LMArena Non-English | 1358 | 1342 |
| LMArena Chinese | 1391 | 1373 |
| LMArena French | 1385 | 1391 |
| LMArena German | 1374 | 1353 |
| LMArena Japanese | 1333 | 1294 |
| LMArena Korean | 1319 | 1313 |
| LMArena Russian | 1373 | 1351 |
| LMArena Spanish | 1358 | 1363 |
Instruction Following Gemini 2.0 Flash (Feb 2025) leads
DeepSeek-V3: 72.8 (#130), Gemini 2.0 Flash (Feb 2025): 74.4 (#97)
| Benchmark | DeepSeek-V3 | Gemini 2.0 Flash (Feb 2025) |
|---|---|---|
| LiveBench Instruction Following | 81.5% | 85.8% |
| IFEval | 83.2% | 84.1% |
| LMArena Instruction Following | 1345 | 1336 |
Long Context Gemini 2.0 Flash (Feb 2025) leads
DeepSeek-V3: 34.0 (#253), Gemini 2.0 Flash (Feb 2025): 38.1 (#203)
| Benchmark | DeepSeek-V3 | Gemini 2.0 Flash (Feb 2025) |
|---|---|---|
| Fiction.LiveBench | 50% | 61.1% |
| LMArena Longer Query | 1352 | 1344 |
Writing & Preference DeepSeek-V3 leads
DeepSeek-V3: 57.4 (#130), Gemini 2.0 Flash (Feb 2025): 49.5 (#190)
| Benchmark | DeepSeek-V3 | Gemini 2.0 Flash (Feb 2025) |
|---|---|---|
| LMArena Text | 1375 | 1354 |
| LMArena Creative Writing | 1364 | 1340 |
| Short-Story Creative Writing | 77% | 73.8% |
| EQ-Bench Creative Writing | 1472 | 1128 |
| WildBench | 83% | 80% |
| LMArena Multi-Turn | 1389 | 1350 |
| LiveBench Language | 49.1% | 51.3% |
Frequently asked questions
Is DeepSeek-V3 better than Gemini 2.0 Flash (Feb 2025)?
DeepSeek-V3 is the stronger model overall, scoring 39.5 to 35.1 on the Noometry Index.
Is DeepSeek-V3 or Gemini 2.0 Flash (Feb 2025) better for coding?
DeepSeek-V3 scores higher on coding benchmarks: 42.3 versus 28.4 in the Noometry coding category.
How many benchmarks do DeepSeek-V3 and Gemini 2.0 Flash (Feb 2025) share?
46 benchmarks have published results for both models. DeepSeek-V3 has 60 scored results on Noometry and Gemini 2.0 Flash (Feb 2025) has 54.