Model comparison
DeepSeek-V3 vs Gemini 3.1 Flash Lite
Gemini 3.1 Flash Lite is the stronger model overall, scoring 40.8 to 39.5 on the Noometry Index.
Last verified . 28 shared benchmarks.
DeepSeek
GoogleSummary
- They share 28 benchmarks with published results for both. DeepSeek-V3 scores higher in 2 categories and Gemini 3.1 Flash Lite in 6 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in math, where Gemini 3.1 Flash Lite leads 40.7 to 32.1.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 37.8% for DeepSeek-V3 and 80% for Gemini 3.1 Flash Lite.
- DeepSeek-V3 is cheaper at $0.24 / $0.90 per million input/output tokens, against $0.25 / $1.50 for Gemini 3.1 Flash Lite.
- Gemini 3.1 Flash Lite accepts more context: 1.05M tokens versus 164K.
- DeepSeek-V3 has downloadable open weights; the other is API-only.
Side by side
| DeepSeek-V3 | Gemini 3.1 Flash Lite | |
|---|---|---|
| Provider | DeepSeek | |
| Noometry Index | 39.5 | 40.8 |
| Released | 2024-12-26 | 2026-03-03 |
| Weights | Open | Proprietary |
| Context window | 164K | 1.05M |
| Max output | 164K | 66K |
| Input $ / M tokens | $0.24 | $0.25 |
| Output $ / M tokens | $0.90 | $1.50 |
| Results tracked | 60 | 38 |
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 3.1 Flash Lite: 37.8 (#188)
| Benchmark | DeepSeek-V3 | Gemini 3.1 Flash Lite |
|---|---|---|
| SciCode | 35.8% | 41.9% |
| WeirdML | 36.1% | 52.2% |
| LMArena Coding | 1368 | 1400 |
| Aider Polyglot | 55.1% | — |
| LMArena WebDev | — | 1256 |
| BigCodeBench Instruct | 50% | — |
| LiveBench Coding | 70.9% | — |
| BigCodeBench Complete | 62.2% | — |
| ALE-Bench | — | 797.73 |
| HumanEval+ | 86.6% | — |
| MBPP+ | 73% | — |
Agentic & Tool Use Not comparable
DeepSeek-V3: —, Gemini 3.1 Flash Lite: 30.2 (#79)
| Benchmark | DeepSeek-V3 | Gemini 3.1 Flash Lite |
|---|---|---|
| DeepResearch Bench | — | 37.3% |
| METR Time Horizons | 49.6% | — |
Reasoning Gemini 3.1 Flash Lite leads
DeepSeek-V3: 20.5 (#236), Gemini 3.1 Flash Lite: 22.9 (#186)
| Benchmark | DeepSeek-V3 | Gemini 3.1 Flash Lite |
|---|---|---|
| Kagi LLM Benchmark | 52.3% | 67.2% |
| CritPt | 0% | 1.1% |
| LMArena Hard Prompts | 1365 | 1407 |
| DTBench | 64.8% | 76.8% |
| LMCA | 15.5% | 35% |
| Epoch Capabilities Index | 135.94 | 144.47 |
| ForecastBench | 59.1 | 54.4 |
| SimpleBench | 27.2% | — |
| NYT Connections (extended) | — | 8.2% |
| Chess Puzzles | — | 25% |
| EnigmaEval | — | 3% |
| Thematic Generalization | — | 63.3% |
| LiveBench Reasoning | 65.8% | — |
| LiveBench Data Analysis | 60.9% | — |
| BIG-Bench Hard | 87.5% | — |
| HellaSwag | 88.9% | — |
| LiveBench | 66.9% | — |
| PIQA | 84.7% | — |
| WinoGrande | 85.2% | — |
Math Gemini 3.1 Flash Lite leads
DeepSeek-V3: 32.1 (#219), Gemini 3.1 Flash Lite: 40.7 (#90)
| Benchmark | DeepSeek-V3 | Gemini 3.1 Flash Lite |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 37.8% | 80% |
| LMArena Math | 1373 | 1428 |
| FrontierMath (Tiers 1-3) | — | 27.7% |
| Omni-MATH | 40.3% | — |
| LiveBench Math | 73.5% | — |
| MATH Level 5 | 75.5% | — |
| FrontierMath (Feb 2025 set) | 1.7% | — |
Knowledge Gemini 3.1 Flash Lite leads
DeepSeek-V3: 37.5 (#155), Gemini 3.1 Flash Lite: 41.9 (#104)
| Benchmark | DeepSeek-V3 | Gemini 3.1 Flash Lite |
|---|---|---|
| GPQA Diamond | 67.6% | 81.8% |
| Vectara Hallucination Rate | 6.1% | 8.2% |
| LMArena Expert | 1351 | 1398 |
| Humanity's Last Exam | — | 8.6% |
| MMLU-Pro | 72.3% | — |
| Confabulations | 26.1% | — |
| GPQA (HELM) | 53.8% | — |
| ARC (AI2) Challenge | 95.3% | — |
| MMLU | 87.2% | — |
| TriviaQA | 82.9% | — |
Multimodal Not comparable
DeepSeek-V3: —, Gemini 3.1 Flash Lite: 39.4 (#60)
| Benchmark | DeepSeek-V3 | Gemini 3.1 Flash Lite |
|---|---|---|
| LMArena Vision | — | 1240 |
Multilingual Gemini 3.1 Flash Lite leads
DeepSeek-V3: 48.5 (#143), Gemini 3.1 Flash Lite: 52.3 (#86)
| Benchmark | DeepSeek-V3 | Gemini 3.1 Flash Lite |
|---|---|---|
| LMArena Non-English | 1358 | 1411 |
| LMArena Chinese | 1391 | 1461 |
| LMArena French | 1385 | 1424 |
| LMArena German | 1374 | 1429 |
| LMArena Japanese | 1333 | 1413 |
| LMArena Korean | 1319 | 1392 |
| LMArena Russian | 1373 | 1420 |
| LMArena Spanish | 1358 | 1421 |
Instruction Following Too close to call
DeepSeek-V3: 72.8 (#130), Gemini 3.1 Flash Lite: 72.7 (#131)
| Benchmark | DeepSeek-V3 | Gemini 3.1 Flash Lite |
|---|---|---|
| LMArena Instruction Following | 1345 | 1377 |
| LiveBench Instruction Following | 81.5% | — |
| IFEval | 83.2% | — |
Long Context Gemini 3.1 Flash Lite leads
DeepSeek-V3: 34.0 (#253), Gemini 3.1 Flash Lite: 42.5 (#122)
| Benchmark | DeepSeek-V3 | Gemini 3.1 Flash Lite |
|---|---|---|
| LMArena Longer Query | 1352 | 1394 |
| Fiction.LiveBench | 50% | — |
Writing & Preference Gemini 3.1 Flash Lite leads
DeepSeek-V3: 57.4 (#130), Gemini 3.1 Flash Lite: 60.9 (#94)
| Benchmark | DeepSeek-V3 | Gemini 3.1 Flash Lite |
|---|---|---|
| LMArena Text | 1375 | 1416 |
| LMArena Creative Writing | 1364 | 1401 |
| LMArena Multi-Turn | 1389 | 1417 |
| Short-Story Creative Writing | 77% | — |
| EQ-Bench Creative Writing | 1472 | — |
| WildBench | 83% | — |
| LiveBench Language | 49.1% | — |
Frequently asked questions
Is DeepSeek-V3 better than Gemini 3.1 Flash Lite?
Gemini 3.1 Flash Lite is the stronger model overall, scoring 40.8 to 39.5 on the Noometry Index.
Which is cheaper, DeepSeek-V3 or Gemini 3.1 Flash Lite?
DeepSeek-V3 is cheaper. It lists at $0.24 per million input tokens and $0.90 per million output tokens; Gemini 3.1 Flash Lite lists at $0.25 and $1.50.
Is DeepSeek-V3 or Gemini 3.1 Flash Lite better for coding?
DeepSeek-V3 scores higher on coding benchmarks: 42.3 versus 37.8 in the Noometry coding category.
Which has the bigger context window?
Gemini 3.1 Flash Lite does, with 1.05M tokens against 164K.
How many benchmarks do DeepSeek-V3 and Gemini 3.1 Flash Lite share?
28 benchmarks have published results for both models. DeepSeek-V3 has 60 scored results on Noometry and Gemini 3.1 Flash Lite has 38.