Model comparison
DeepSeek-V2.5 (Sep 2024) vs Gemini 3.1 Pro Preview
Gemini 3.1 Pro Preview is the stronger model overall, scoring 56.7 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 3.1 Pro Preview in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Gemini 3.1 Pro Preview leads 71.7 to 25.6.
- DeepSeek-V2.5 (Sep 2024) has downloadable open weights; the other is API-only.
Side by side
| DeepSeek-V2.5 (Sep 2024) | Gemini 3.1 Pro Preview | |
|---|---|---|
| Provider | DeepSeek | |
| Noometry Index | 37.6 | 56.7 |
| Released | 2024-09-06 | 2026-02-19 |
| Weights | Open | Proprietary |
| Context window | — | 1.05M |
| Max output | — | 66K |
| Input $ / M tokens | — | $2 |
| Output $ / M tokens | — | $12 |
| Results tracked | 22 | 71 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Gemini 3.1 Pro Preview leads
DeepSeek-V2.5 (Sep 2024): 31.7 (#281), Gemini 3.1 Pro Preview: 42.5 (#99)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Gemini 3.1 Pro Preview |
|---|---|---|
| LMArena Coding | 1309 | 1484 |
| SWE-bench Verified | — | 75.6% |
| DeepSWE | — | 11.7% |
| Aider Polyglot | 17.8% | — |
| LMArena WebDev | — | 1447 |
| SciCode | — | 58.9% |
| GSO | — | 22.6% |
| WeirdML | — | 72.1% |
| BigCodeBench Instruct | 48.6% | — |
| MirrorCode | — | 8.9% |
| BigCodeBench Complete | 53.2% | — |
| ALE-Bench | — | 1,161 |
| AlgoTune | — | 2.02 |
| HumanEval+ | 83.5% | — |
| MBPP+ | 74.1% | — |
Agentic & Tool Use Not comparable
DeepSeek-V2.5 (Sep 2024): —, Gemini 3.1 Pro Preview: 37.7 (#34)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Gemini 3.1 Pro Preview |
|---|---|---|
| Terminal-Bench | — | 80.2% |
| APEX-Agents | — | 35.3% |
| τ²-bench Banking | — | 26% |
| DeepResearch Bench | — | 47.8% |
| PostTrainBench | — | 22% |
| BALROG | — | 57% |
| ExploitBench | — | 26.1% |
| GBAEval | — | 0.8% |
| GDP.pdf | — | 17% |
| LMArena Search | — | 1211 |
| METR Time Horizons | — | 77% |
| Vending-Bench 2 | — | 3,774 |
Reasoning Gemini 3.1 Pro Preview leads
DeepSeek-V2.5 (Sep 2024): 25.6 (#145), Gemini 3.1 Pro Preview: 71.7 (#12)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Gemini 3.1 Pro Preview |
|---|---|---|
| LMArena Hard Prompts | 1289 | 1485 |
| ARC-AGI-2 | — | 77.1% |
| SimpleBench | — | 79.6% |
| NYT Connections (extended) | — | 97.4% |
| ARC-AGI-1 | — | 98% |
| CritPt | — | 17.7% |
| Chess Puzzles | — | 55% |
| EnigmaEval | — | 36.8% |
| Thematic Generalization | — | 79.4% |
| EBR-Bench | — | 14.3% |
| Mystery Game Puzzles | — | 34% |
| DTBench | — | 97.1% |
| LMCA | — | 53.8% |
| Epoch Capabilities Index | — | 154.77 |
| ForecastBench | — | 59 |
Math Gemini 3.1 Pro Preview leads
DeepSeek-V2.5 (Sep 2024): 35.9 (#177), Gemini 3.1 Pro Preview: 62.1 (#34)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Gemini 3.1 Pro Preview |
|---|---|---|
| LMArena Math | 1288 | 1485 |
| FrontierMath (Tiers 1-3) | — | 59.6% |
| FrontierMath Tier 4 | — | 26.8% |
| MathArena Final-Answer Competitions | — | 86.5% |
| OTIS Mock AIME 2024-2025 | — | 95.6% |
| ProofBench | — | 26% |
| FrontierMath (Feb 2025 set) | — | 36.9% |
| FrontierMath Tier 4 (v1) | — | 16.7% |
Knowledge Gemini 3.1 Pro Preview leads
DeepSeek-V2.5 (Sep 2024): 34.8 (#193), Gemini 3.1 Pro Preview: 71.8 (#3)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Gemini 3.1 Pro Preview |
|---|---|---|
| LMArena Expert | 1266 | 1485 |
| GPQA Diamond | — | 94.4% |
| Humanity's Last Exam | — | 46.4% |
| SimpleQA Verified | — | 73.5% |
| Vectara Hallucination Rate | — | 10.4% |
Multimodal Not comparable
DeepSeek-V2.5 (Sep 2024): —, Gemini 3.1 Pro Preview: 37.9 (#69)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Gemini 3.1 Pro Preview |
|---|---|---|
| LMArena Vision | — | 1296 |
| Blueprint-Bench 2 | — | 26.5% |
| Furniture Assembly | — | 26.7% |
| LMArena Document | — | 1444 |
Multilingual Gemini 3.1 Pro Preview leads
DeepSeek-V2.5 (Sep 2024): 42.5 (#193), Gemini 3.1 Pro Preview: 57.0 (#12)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Gemini 3.1 Pro Preview |
|---|---|---|
| LMArena Non-English | 1273 | 1477 |
| LMArena Chinese | 1318 | 1529 |
| LMArena French | 1289 | 1487 |
| LMArena German | 1258 | 1491 |
| LMArena Japanese | 1228 | 1493 |
| LMArena Korean | 1209 | 1455 |
| LMArena Russian | 1289 | 1498 |
| LMArena Spanish | 1248 | 1479 |
Instruction Following Gemini 3.1 Pro Preview leads
DeepSeek-V2.5 (Sep 2024): 67.5 (#194), Gemini 3.1 Pro Preview: 77.0 (#32)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Gemini 3.1 Pro Preview |
|---|---|---|
| LMArena Instruction Following | 1280 | 1466 |
Long Context Gemini 3.1 Pro Preview leads
DeepSeek-V2.5 (Sep 2024): 39.5 (#174), Gemini 3.1 Pro Preview: 47.4 (#18)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Gemini 3.1 Pro Preview |
|---|---|---|
| LMArena Longer Query | 1301 | 1483 |
| CL-bench | — | 20.8% |
| CL-bench Life | — | 16.9% |
Writing & Preference Gemini 3.1 Pro Preview leads
DeepSeek-V2.5 (Sep 2024): 49.8 (#187), Gemini 3.1 Pro Preview: 66.1 (#37)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | Gemini 3.1 Pro Preview |
|---|---|---|
| LMArena Text | 1294 | 1481 |
| LMArena Creative Writing | 1285 | 1482 |
| LMArena Multi-Turn | 1297 | 1488 |
| EQ-Bench Creative Writing | — | 1491 |
| EQ-Bench 4 | — | 1142 |
Frequently asked questions
Is DeepSeek-V2.5 (Sep 2024) better than Gemini 3.1 Pro Preview?
Gemini 3.1 Pro Preview is the stronger model overall, scoring 56.7 to 37.6 on the Noometry Index.
Is DeepSeek-V2.5 (Sep 2024) or Gemini 3.1 Pro Preview better for coding?
Gemini 3.1 Pro Preview scores higher on coding benchmarks: 42.5 versus 31.7 in the Noometry coding category.
How many benchmarks do DeepSeek-V2.5 (Sep 2024) and Gemini 3.1 Pro Preview share?
17 benchmarks have published results for both models. DeepSeek-V2.5 (Sep 2024) has 22 scored results on Noometry and Gemini 3.1 Pro Preview has 71.