Model comparison
DeepSeek V4 Pro vs Gemini 3.5 Flash Lite
DeepSeek V4 Pro is the stronger model overall, scoring 54.2 to 41.8 on the Noometry Index.
Last verified . 36 shared benchmarks.
DeepSeek
GoogleSummary
- They share 36 benchmarks with published results for both. DeepSeek V4 Pro scores higher in 9 categories and Gemini 3.5 Flash Lite in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where DeepSeek V4 Pro leads 64.6 to 25.7.
- The biggest single-benchmark swing is ARC-AGI-2: 61.3% for DeepSeek V4 Pro and 10.3% for Gemini 3.5 Flash Lite.
- DeepSeek V4 Pro is cheaper at $0.43 / $0.87 per million input/output tokens, against $0.30 / $2.50 for Gemini 3.5 Flash Lite.
- Gemini 3.5 Flash Lite 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 3.5 Flash Lite | |
|---|---|---|
| Provider | DeepSeek | |
| Noometry Index | 54.2 | 41.8 |
| Released | 2026-04-24 | 2026-07-21 |
| Weights | Open | Proprietary |
| Context window | 1M | 1.05M |
| Max output | 393K | 66K |
| Input $ / M tokens | $0.43 | $0.30 |
| Output $ / M tokens | $0.87 | $2.50 |
| Results tracked | 48 | 38 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek V4 Pro leads
DeepSeek V4 Pro: 52.2 (#31), Gemini 3.5 Flash Lite: 42.1 (#96)
| Benchmark | DeepSeek V4 Pro | Gemini 3.5 Flash Lite |
|---|---|---|
| LMArena WebDev | 1582 | 1440 |
| SciCode | 51% | 41.3% |
| WeirdML | 66.2% | 39% |
| LMArena Coding | 1470 | 1455 |
| ALE-Bench | 1,403 | 765.27 |
| SWE-bench Verified | 77.6% | — |
| FrontierCode | 28.6% | — |
Agentic & Tool Use DeepSeek V4 Pro leads
DeepSeek V4 Pro: 32.7 (#54), Gemini 3.5 Flash Lite: 28.6 (#82)
| Benchmark | DeepSeek V4 Pro | Gemini 3.5 Flash Lite |
|---|---|---|
| APEX-Agents | 47.3% | — |
| GDP.pdf | — | 10% |
| Vending-Bench 2 | 3,285 | — |
Reasoning DeepSeek V4 Pro leads
DeepSeek V4 Pro: 56.4 (#22), Gemini 3.5 Flash Lite: 27.8 (#95)
| Benchmark | DeepSeek V4 Pro | Gemini 3.5 Flash Lite |
|---|---|---|
| ARC-AGI-2 | 61.3% | 10.3% |
| NYT Connections (extended) | 91.3% | 60.4% |
| ARC-AGI-1 | 90.5% | 53.5% |
| CritPt | 18% | 0% |
| Chess Puzzles | 47% | 22% |
| LMArena Hard Prompts | 1461 | 1441 |
| Mystery Game Puzzles | 43% | 19% |
| DTBench | 93.9% | 83.5% |
| LMCA | 45.5% | 37.6% |
| Epoch Capabilities Index | 155.39 | 145.12 |
| Kagi LLM Benchmark | 53.5% | — |
| Surface Evolver Bench | 40% | — |
| ForecastBench | 56.1 | — |
Math DeepSeek V4 Pro leads
DeepSeek V4 Pro: 64.6 (#24), Gemini 3.5 Flash Lite: 25.7 (#245)
| Benchmark | DeepSeek V4 Pro | Gemini 3.5 Flash Lite |
|---|---|---|
| FrontierMath (Tiers 1-3) | 64.6% | 26% |
| FrontierMath Tier 4 | 26.8% | 0% |
| OTIS Mock AIME 2024-2025 | 98.6% | 71.1% |
| ProofBench | 50% | 13% |
| LMArena Math | 1455 | 1431 |
| MathArena Final-Answer Competitions | 76.6% | — |
Knowledge DeepSeek V4 Pro leads
DeepSeek V4 Pro: 59.5 (#28), Gemini 3.5 Flash Lite: 50.0 (#66)
| Benchmark | DeepSeek V4 Pro | Gemini 3.5 Flash Lite |
|---|---|---|
| GPQA Diamond | 91.7% | 83.3% |
| LMArena Expert | 1463 | 1430 |
| SimpleQA Verified | 52.9% | — |
| Vectara Hallucination Rate | 8.6% | — |
Multimodal Not comparable
DeepSeek V4 Pro: —, Gemini 3.5 Flash Lite: 40.3 (#38)
| Benchmark | DeepSeek V4 Pro | Gemini 3.5 Flash Lite |
|---|---|---|
| LMArena Vision | — | 1266 |
Multilingual Too close to call
DeepSeek V4 Pro: 54.5 (#44), Gemini 3.5 Flash Lite: 53.8 (#59)
| Benchmark | DeepSeek V4 Pro | Gemini 3.5 Flash Lite |
|---|---|---|
| LMArena Non-English | 1438 | 1428 |
| LMArena Chinese | 1486 | 1467 |
| LMArena French | 1473 | 1452 |
| LMArena German | 1458 | 1454 |
| LMArena Japanese | 1448 | 1428 |
| LMArena Korean | 1452 | 1406 |
| LMArena Russian | 1452 | 1443 |
| LMArena Spanish | 1457 | 1440 |
Instruction Following DeepSeek V4 Pro leads
DeepSeek V4 Pro: 76.3 (#44), Gemini 3.5 Flash Lite: 75.0 (#77)
| Benchmark | DeepSeek V4 Pro | Gemini 3.5 Flash Lite |
|---|---|---|
| LMArena Instruction Following | 1448 | 1420 |
Long Context DeepSeek V4 Pro leads
DeepSeek V4 Pro: 45.0 (#49), Gemini 3.5 Flash Lite: 43.8 (#80)
| Benchmark | DeepSeek V4 Pro | Gemini 3.5 Flash Lite |
|---|---|---|
| LMArena Longer Query | 1459 | 1435 |
| CL-bench Life | 13.5% | — |
Writing & Preference DeepSeek V4 Pro leads
DeepSeek V4 Pro: 65.6 (#45), Gemini 3.5 Flash Lite: 64.4 (#55)
| Benchmark | DeepSeek V4 Pro | Gemini 3.5 Flash Lite |
|---|---|---|
| LMArena Text | 1451 | 1436 |
| LMArena Creative Writing | 1446 | 1419 |
| EQ-Bench Creative Writing | 1553 | 1559 |
| LMArena Multi-Turn | 1467 | 1445 |
| EQ-Bench 4 | 1166 | — |
Frequently asked questions
Is DeepSeek V4 Pro better than Gemini 3.5 Flash Lite?
DeepSeek V4 Pro is the stronger model overall, scoring 54.2 to 41.8 on the Noometry Index.
Which is cheaper, DeepSeek V4 Pro or Gemini 3.5 Flash Lite?
DeepSeek V4 Pro is cheaper. It lists at $0.43 per million input tokens and $0.87 per million output tokens; Gemini 3.5 Flash Lite lists at $0.30 and $2.50.
Is DeepSeek V4 Pro or Gemini 3.5 Flash Lite better for coding?
DeepSeek V4 Pro scores higher on coding benchmarks: 52.2 versus 42.1 in the Noometry coding category.
Which has the bigger context window?
Gemini 3.5 Flash Lite does, with 1.05M tokens against 1M.
How many benchmarks do DeepSeek V4 Pro and Gemini 3.5 Flash Lite share?
36 benchmarks have published results for both models. DeepSeek V4 Pro has 48 scored results on Noometry and Gemini 3.5 Flash Lite has 38.