Model comparison
DeepSeek-V2.5 (Sep 2024) vs Gemini 3 Flash Preview
Gemini 3 Flash Preview is the stronger model overall, scoring 52.3 to 37.6 on the Noometry Index.
Last verified . 17 shared benchmarks.
DeepSeek
GoogleSummary
- They share 17 benchmarks with published results for both. DeepSeek-V2.5 (Sep 2024) scores higher in 0 categories and Gemini 3 Flash Preview in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Gemini 3 Flash Preview leads 58.8 to 34.8.
- DeepSeek-V2.5 (Sep 2024) has downloadable open weights; the other is API-only.
Side by side
| DeepSeek-V2.5 (Sep 2024) | Gemini 3 Flash Preview | |
|---|---|---|
| Provider | DeepSeek | |
| Noometry Index | 37.6 | 52.3 |
| Released | 2024-09-06 | 2025-12-17 |
| Weights | Open | Proprietary |
| Context window | — | 1.05M |
| Max output | — | 66K |
| Input $ / M tokens | — | $0.50 |
| Output $ / M tokens | — | $3 |
| Results tracked | 22 | 59 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Gemini 3 Flash Preview leads
DeepSeek-V2.5 (Sep 2024): 31.7 (#281), Gemini 3 Flash Preview: 50.9 (#42)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Gemini 3 Flash Preview |
|---|---|---|
| LMArena Coding | 1309 | 1460 |
| SWE-bench Verified | — | 75.4% |
| SWE-bench Verified (bash only) | — | 75.8% |
| Aider Polyglot | 17.8% | — |
| LMArena WebDev | — | 1439 |
| SWE-bench Multilingual | — | 72.7% |
| GSO | — | 9.8% |
| WeirdML | — | 61.6% |
| BigCodeBench Instruct | 48.6% | — |
| BigCodeBench Complete | 53.2% | — |
| ALE-Bench | — | 1,367 |
| HumanEval+ | 83.5% | — |
| MBPP+ | 74.1% | — |
Agentic & Tool Use Not comparable
DeepSeek-V2.5 (Sep 2024): —, Gemini 3 Flash Preview: 38.7 (#29)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Gemini 3 Flash Preview |
|---|---|---|
| Terminal-Bench | — | 64.3% |
| τ²-bench Airline | — | 82.5% |
| τ²-bench Banking | — | 27.3% |
| τ²-bench Retail | — | 76.8% |
| τ²-bench Telecom | — | 91.2% |
| DeepResearch Bench | — | 49.8% |
| BALROG | — | 48.1% |
| GDP.pdf | — | 10% |
| LMArena Search | — | 1198 |
| Vending-Bench 2 | — | 3,635 |
Reasoning Gemini 3 Flash Preview leads
DeepSeek-V2.5 (Sep 2024): 25.6 (#145), Gemini 3 Flash Preview: 49.2 (#37)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Gemini 3 Flash Preview |
|---|---|---|
| LMArena Hard Prompts | 1289 | 1465 |
| ARC-AGI-2 | — | 33.6% |
| SimpleBench | — | 61.1% |
| NYT Connections (extended) | — | 83.1% |
| ARC-AGI-1 | — | 84.7% |
| Chess Puzzles | — | 40% |
| Mystery Game Puzzles | — | 26% |
| DTBench | — | 89.1% |
| LMCA | — | 43.1% |
| Epoch Capabilities Index | — | 151.8 |
| ForecastBench | — | 58.5 |
Math Gemini 3 Flash Preview leads
DeepSeek-V2.5 (Sep 2024): 35.9 (#177), Gemini 3 Flash Preview: 51.7 (#55)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Gemini 3 Flash Preview |
|---|---|---|
| LMArena Math | 1288 | 1473 |
| FrontierMath (Tiers 1-3) | — | 51.2% |
| FrontierMath Tier 4 | — | 17.1% |
| MathArena Final-Answer Competitions | — | 67.6% |
| OTIS Mock AIME 2024-2025 | — | 95.6% |
| ProofBench | — | 15% |
| FrontierMath (Feb 2025 set) | — | 35.6% |
| FrontierMath Tier 4 (v1) | — | 4.2% |
Knowledge Gemini 3 Flash Preview leads
DeepSeek-V2.5 (Sep 2024): 34.8 (#193), Gemini 3 Flash Preview: 58.8 (#33)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Gemini 3 Flash Preview |
|---|---|---|
| LMArena Expert | 1266 | 1462 |
| GPQA Diamond | — | 89.4% |
| SimpleQA Verified | — | 66.8% |
| Vectara Hallucination Rate | — | 13.5% |
Multimodal Not comparable
DeepSeek-V2.5 (Sep 2024): —, Gemini 3 Flash Preview: 45.5 (#16)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Gemini 3 Flash Preview |
|---|---|---|
| LMArena Vision | — | 1285 |
| GeoBench | — | 88% |
| VPCT | — | 72.6% |
| Blueprint-Bench 2 | — | 0% |
| LMArena Document | — | 1413 |
Multilingual Gemini 3 Flash Preview leads
DeepSeek-V2.5 (Sep 2024): 42.5 (#193), Gemini 3 Flash Preview: 55.7 (#27)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Gemini 3 Flash Preview |
|---|---|---|
| LMArena Non-English | 1273 | 1458 |
| LMArena Chinese | 1318 | 1511 |
| LMArena French | 1289 | 1477 |
| LMArena German | 1258 | 1497 |
| LMArena Japanese | 1228 | 1489 |
| LMArena Korean | 1209 | 1443 |
| LMArena Russian | 1289 | 1480 |
| LMArena Spanish | 1248 | 1469 |
Instruction Following Gemini 3 Flash Preview leads
DeepSeek-V2.5 (Sep 2024): 67.5 (#194), Gemini 3 Flash Preview: 75.7 (#56)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Gemini 3 Flash Preview |
|---|---|---|
| LMArena Instruction Following | 1280 | 1437 |
Long Context Gemini 3 Flash Preview leads
DeepSeek-V2.5 (Sep 2024): 39.5 (#174), Gemini 3 Flash Preview: 44.4 (#67)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Gemini 3 Flash Preview |
|---|---|---|
| LMArena Longer Query | 1301 | 1452 |
Writing & Preference Gemini 3 Flash Preview leads
DeepSeek-V2.5 (Sep 2024): 49.8 (#187), Gemini 3 Flash Preview: 65.5 (#45)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Gemini 3 Flash Preview |
|---|---|---|
| LMArena Text | 1294 | 1466 |
| LMArena Creative Writing | 1285 | 1457 |
| LMArena Multi-Turn | 1297 | 1471 |
Frequently asked questions
Is DeepSeek-V2.5 (Sep 2024) better than Gemini 3 Flash Preview?
Gemini 3 Flash Preview is the stronger model overall, scoring 52.3 to 37.6 on the Noometry Index.
Is DeepSeek-V2.5 (Sep 2024) or Gemini 3 Flash Preview better for coding?
Gemini 3 Flash Preview scores higher on coding benchmarks: 50.9 versus 31.7 in the Noometry coding category.
How many benchmarks do DeepSeek-V2.5 (Sep 2024) and Gemini 3 Flash Preview share?
17 benchmarks have published results for both models. DeepSeek-V2.5 (Sep 2024) has 22 scored results on Noometry and Gemini 3 Flash Preview has 59.