Model comparison
DeepSeek V4 Pro vs Gemini 3.1 Pro Preview
Gemini 3.1 Pro Preview is the stronger model overall, scoring 56.7 to 54.3 on the Noometry Index. DeepSeek V4 Pro costs 4.5× less per token, which makes it the better buy when Gemini 3.1 Pro Preview's lead doesn't matter for your workload.
Last verified . 45 shared benchmarks.
DeepSeek
GoogleSummary
- They share 45 benchmarks with published results for both. DeepSeek V4 Pro scores higher in 2 categories and Gemini 3.1 Pro Preview in 7 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Gemini 3.1 Pro Preview leads 71.7 to 56.5.
- The biggest single-benchmark swing is ProofBench: 50% for DeepSeek V4 Pro and 26% for Gemini 3.1 Pro Preview.
- DeepSeek V4 Pro is cheaper at $0.66 / $1.98 per million input/output tokens, against $2 / $12 for Gemini 3.1 Pro Preview.
- Gemini 3.1 Pro Preview accepts more context: 1.05M tokens versus 1M.
- DeepSeek V4 Pro has downloadable open weights; the other is API-only.
Side by side
| DeepSeek V4 Pro | Gemini 3.1 Pro Preview | |
|---|---|---|
| Provider | DeepSeek | |
| Noometry Index | 54.3 | 56.7 |
| Released | 2026-04-24 | 2026-02-19 |
| Weights | Open | Proprietary |
| Context window | 1M | 1.05M |
| Max output | 393K | 66K |
| Input $ / M tokens | $0.66 | $2 |
| Output $ / M tokens | $1.98 | $12 |
| Results tracked | 48 | 71 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek V4 Pro leads
DeepSeek V4 Pro: 52.4 (#34), Gemini 3.1 Pro Preview: 42.5 (#99)
| Benchmark | DeepSeek V4 Pro | Gemini 3.1 Pro Preview |
|---|---|---|
| SWE-bench Verified | 77.6% | 75.6% |
| LMArena WebDev | 1582 | 1447 |
| SciCode | 51% | 58.9% |
| WeirdML | 66.2% | 72.1% |
| LMArena Coding | 1470 | 1484 |
| ALE-Bench | 1,403 | 1,161 |
| DeepSWE | — | 11.7% |
| FrontierCode | 28.6% | — |
| GSO | — | 22.6% |
| MirrorCode | — | 8.9% |
| AlgoTune | — | 2.02 |
Agentic & Tool Use Gemini 3.1 Pro Preview leads
DeepSeek V4 Pro: 32.8 (#58), Gemini 3.1 Pro Preview: 37.7 (#34)
| Benchmark | DeepSeek V4 Pro | Gemini 3.1 Pro Preview |
|---|---|---|
| APEX-Agents | 47.3% | 35.3% |
| Vending-Bench 2 | 3,285 | 3,774 |
| Terminal-Bench | — | 80.2% |
| 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% |
Reasoning Gemini 3.1 Pro Preview leads
DeepSeek V4 Pro: 56.5 (#24), Gemini 3.1 Pro Preview: 71.7 (#12)
| Benchmark | DeepSeek V4 Pro | Gemini 3.1 Pro Preview |
|---|---|---|
| ARC-AGI-2 | 61.3% | 77.1% |
| NYT Connections (extended) | 91.3% | 97.4% |
| ARC-AGI-1 | 90.5% | 98% |
| CritPt | 18% | 17.7% |
| Chess Puzzles | 47% | 55% |
| LMArena Hard Prompts | 1461 | 1485 |
| Mystery Game Puzzles | 43% | 34% |
| DTBench | 93.9% | 97.1% |
| LMCA | 45.5% | 53.8% |
| Epoch Capabilities Index | 155.31 | 154.77 |
| ForecastBench | 56.1 | 59 |
| SimpleBench | — | 79.6% |
| Kagi LLM Benchmark | 53.5% | — |
| EnigmaEval | — | 36.8% |
| Thematic Generalization | — | 79.4% |
| EBR-Bench | — | 14.3% |
| Surface Evolver Bench | 40% | — |
Math DeepSeek V4 Pro leads
DeepSeek V4 Pro: 64.8 (#30), Gemini 3.1 Pro Preview: 62.1 (#34)
| Benchmark | DeepSeek V4 Pro | Gemini 3.1 Pro Preview |
|---|---|---|
| FrontierMath (Tiers 1-3) | 64.6% | 59.6% |
| FrontierMath Tier 4 | 26.8% | 26.8% |
| MathArena Final-Answer Competitions | 76.6% | 86.5% |
| OTIS Mock AIME 2024-2025 | 98.6% | 95.6% |
| ProofBench | 50% | 26% |
| LMArena Math | 1455 | 1485 |
| FrontierMath (Feb 2025 set) | — | 36.9% |
| FrontierMath Tier 4 (v1) | — | 16.7% |
Knowledge Gemini 3.1 Pro Preview leads
DeepSeek V4 Pro: 59.5 (#31), Gemini 3.1 Pro Preview: 71.8 (#3)
| Benchmark | DeepSeek V4 Pro | Gemini 3.1 Pro Preview |
|---|---|---|
| GPQA Diamond | 91.7% | 94.4% |
| SimpleQA Verified | 52.9% | 73.5% |
| Vectara Hallucination Rate | 8.6% | 10.4% |
| LMArena Expert | 1464 | 1485 |
| Humanity's Last Exam | — | 46.4% |
Multimodal Not comparable
DeepSeek V4 Pro: —, Gemini 3.1 Pro Preview: 37.9 (#69)
| Benchmark | DeepSeek V4 Pro | 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 V4 Pro: 54.4 (#45), Gemini 3.1 Pro Preview: 57.0 (#12)
| Benchmark | DeepSeek V4 Pro | Gemini 3.1 Pro Preview |
|---|---|---|
| LMArena Non-English | 1439 | 1477 |
| LMArena Chinese | 1486 | 1529 |
| LMArena French | 1472 | 1487 |
| LMArena German | 1458 | 1491 |
| LMArena Japanese | 1445 | 1493 |
| LMArena Korean | 1447 | 1455 |
| LMArena Russian | 1453 | 1498 |
| LMArena Spanish | 1458 | 1479 |
Instruction Following Too close to call
DeepSeek V4 Pro: 76.1 (#47), Gemini 3.1 Pro Preview: 77.0 (#32)
| Benchmark | DeepSeek V4 Pro | Gemini 3.1 Pro Preview |
|---|---|---|
| LMArena Instruction Following | 1448 | 1466 |
Long Context Gemini 3.1 Pro Preview leads
DeepSeek V4 Pro: 45.0 (#51), Gemini 3.1 Pro Preview: 47.4 (#18)
| Benchmark | DeepSeek V4 Pro | Gemini 3.1 Pro Preview |
|---|---|---|
| CL-bench Life | 13.5% | 16.9% |
| LMArena Longer Query | 1458 | 1483 |
| CL-bench | — | 20.8% |
Writing & Preference Too close to call
DeepSeek V4 Pro: 65.5 (#46), Gemini 3.1 Pro Preview: 66.1 (#37)
| Benchmark | DeepSeek V4 Pro | Gemini 3.1 Pro Preview |
|---|---|---|
| LMArena Text | 1451 | 1481 |
| LMArena Creative Writing | 1446 | 1482 |
| EQ-Bench Creative Writing | 1553 | 1491 |
| EQ-Bench 4 | 1166 | 1142 |
| LMArena Multi-Turn | 1467 | 1488 |
Frequently asked questions
Is DeepSeek V4 Pro better than Gemini 3.1 Pro Preview?
Gemini 3.1 Pro Preview is the stronger model overall, scoring 56.7 to 54.3 on the Noometry Index. DeepSeek V4 Pro costs 4.5× less per token, which makes it the better buy when Gemini 3.1 Pro Preview's lead doesn't matter for your workload.
Which is cheaper, DeepSeek V4 Pro or Gemini 3.1 Pro Preview?
DeepSeek V4 Pro is cheaper. It lists at $0.66 per million input tokens and $1.98 per million output tokens; Gemini 3.1 Pro Preview lists at $2 and $12.
Is DeepSeek V4 Pro or Gemini 3.1 Pro Preview better for coding?
DeepSeek V4 Pro scores higher on coding benchmarks: 52.4 versus 42.5 in the Noometry coding category.
Which has the bigger context window?
Gemini 3.1 Pro Preview does, with 1.05M tokens against 1M.
How many benchmarks do DeepSeek V4 Pro and Gemini 3.1 Pro Preview share?
45 benchmarks have published results for both models. DeepSeek V4 Pro has 48 scored results on Noometry and Gemini 3.1 Pro Preview has 71.