Model comparison
DeepSeek-R1 vs Gemini 3.7 Flash
Gemini 3.7 Flash is the stronger model overall, scoring 59.8 to 42.3 on the Noometry Index. DeepSeek-R1 costs 1.6× less per token, which makes it the better buy when Gemini 3.7 Flash's lead doesn't matter for your workload.
Last verified . 26 shared benchmarks.
DeepSeek
GoogleSummary
- They share 26 benchmarks with published results for both. DeepSeek-R1 scores higher in 0 categories and Gemini 3.7 Flash in 9 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Gemini 3.7 Flash leads 70.0 to 18.6.
- The biggest single-benchmark swing is ARC-AGI-2: 1.3% for DeepSeek-R1 and 84.6% for Gemini 3.7 Flash.
- DeepSeek-R1 is cheaper at $0.50 / $2.15 per million input/output tokens, against $0.75 / $3.75 for Gemini 3.7 Flash.
- Gemini 3.7 Flash accepts more context: 1.05M tokens versus 164K.
Side by side
| DeepSeek-R1 | Gemini 3.7 Flash | |
|---|---|---|
| Provider | DeepSeek | |
| Noometry Index | 42.3 | 59.8 |
| Released | 2025-01-20 | 2026-08-13 |
| Weights | Proprietary | Proprietary |
| Context window | 164K | 1.05M |
| Max output | 64K | 66K |
| Input $ / M tokens | $0.50 | $0.75 |
| Output $ / M tokens | $2.15 | $3.75 |
| Results tracked | 52 | 44 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Gemini 3.7 Flash leads
DeepSeek-R1: 46.3 (#68), Gemini 3.7 Flash: 56.2 (#22)
| Benchmark | DeepSeek-R1 | Gemini 3.7 Flash |
|---|---|---|
| SciCode | 35.7% | 59.8% |
| LMArena Coding | 1427 | 1497 |
| ALE-Bench | 804.12 | 904.3 |
| DeepSWE | — | 65.5% |
| FrontierCode | — | 43.6% |
| Aider Polyglot | 71.4% | — |
| LMArena WebDev | — | 1592 |
| FrontierSWE | — | 20.3% |
| WeirdML | 41.6% | — |
| LiveBench Coding | 66.7% | — |
| AlgoTune | 1.7 | — |
Agentic & Tool Use Gemini 3.7 Flash leads
DeepSeek-R1: 30.7 (#75), Gemini 3.7 Flash: 42.1 (#19)
| Benchmark | DeepSeek-R1 | Gemini 3.7 Flash |
|---|---|---|
| APEX-Agents | — | 67.8% |
| Remote Labor Index | — | 5% |
| DeepResearch Bench | 35.1% | — |
| BALROG | 34.9% | — |
| GDP.pdf | — | 23.8% |
| METR Time Horizons | 53.8% | — |
Reasoning Gemini 3.7 Flash leads
DeepSeek-R1: 18.6 (#278), Gemini 3.7 Flash: 70.0 (#15)
| Benchmark | DeepSeek-R1 | Gemini 3.7 Flash |
|---|---|---|
| ARC-AGI-2 | 1.3% | 84.6% |
| ARC-AGI-1 | 21.2% | 95.5% |
| CritPt | 1.1% | 14.3% |
| LMArena Hard Prompts | 1416 | 1494 |
| Epoch Capabilities Index | 141.29 | 157.27 |
| SimpleBench | 40.8% | — |
| Kagi LLM Benchmark | 69.4% | — |
| NYT Connections (extended) | — | 94% |
| Chess Puzzles | — | 47% |
| LiveBench Reasoning | 83.2% | — |
| Mystery Game Puzzles | — | 37% |
| DTBench | — | 96.8% |
| LiveBench Data Analysis | 69.8% | — |
| LMCA | — | 50.4% |
| ForecastBench | 60 | — |
| LiveBench | 71.6% | — |
Math Gemini 3.7 Flash leads
DeepSeek-R1: 43.8 (#79), Gemini 3.7 Flash: 69.6 (#23)
| Benchmark | DeepSeek-R1 | Gemini 3.7 Flash |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 66.4% | 97.2% |
| LMArena Math | 1400 | 1507 |
| FrontierMath (Tiers 1-3) | — | 71.6% |
| FrontierMath Tier 4 | — | 36.6% |
| ProofBench | — | 58% |
| Omni-MATH | 42.4% | — |
| LiveBench Math | 80.7% | — |
| MATH Level 5 | 96.6% | — |
Knowledge Gemini 3.7 Flash leads
DeepSeek-R1: 44.5 (#87), Gemini 3.7 Flash: 69.7 (#5)
| Benchmark | DeepSeek-R1 | Gemini 3.7 Flash |
|---|---|---|
| GPQA Diamond | 76.3% | 94.8% |
| LMArena Expert | 1394 | 1508 |
| SimpleQA Verified | — | 69.2% |
| MMLU-Pro | 79.3% | — |
| Confabulations | 12.7% | — |
| Vectara Hallucination Rate | 11.3% | — |
| GPQA (HELM) | 66.6% | — |
Multimodal Not comparable
DeepSeek-R1: —, Gemini 3.7 Flash: 37.3 (#73)
| Benchmark | DeepSeek-R1 | Gemini 3.7 Flash |
|---|---|---|
| LMArena Vision | — | 1316 |
| Furniture Assembly | — | 26.7% |
Multilingual Gemini 3.7 Flash leads
DeepSeek-R1: 52.4 (#85), Gemini 3.7 Flash: 57.6 (#7)
| Benchmark | DeepSeek-R1 | Gemini 3.7 Flash |
|---|---|---|
| LMArena Non-English | 1412 | 1484 |
| LMArena Chinese | 1442 | 1548 |
| LMArena French | 1417 | 1505 |
| LMArena German | 1404 | 1498 |
| LMArena Japanese | 1391 | 1512 |
| LMArena Korean | 1360 | 1483 |
| LMArena Russian | 1423 | 1516 |
| LMArena Spanish | 1411 | 1503 |
Instruction Following Gemini 3.7 Flash leads
DeepSeek-R1: 72.0 (#143), Gemini 3.7 Flash: 77.7 (#15)
| Benchmark | DeepSeek-R1 | Gemini 3.7 Flash |
|---|---|---|
| LMArena Instruction Following | 1382 | 1483 |
| LiveBench Instruction Following | 80.5% | — |
| IFEval | 78.4% | — |
Long Context Too close to call
DeepSeek-R1: 45.4 (#36), Gemini 3.7 Flash: 45.7 (#30)
| Benchmark | DeepSeek-R1 | Gemini 3.7 Flash |
|---|---|---|
| LMArena Longer Query | 1391 | 1492 |
| Fiction.LiveBench | 75% | — |
Writing & Preference Gemini 3.7 Flash leads
DeepSeek-R1: 61.4 (#88), Gemini 3.7 Flash: 71.2 (#20)
| Benchmark | DeepSeek-R1 | Gemini 3.7 Flash |
|---|---|---|
| LMArena Text | 1428 | 1486 |
| LMArena Creative Writing | 1405 | 1490 |
| EQ-Bench Creative Writing | 1500 | 1723 |
| LMArena Multi-Turn | 1405 | 1489 |
| Short-Story Creative Writing | 83% | — |
| WildBench | 82.8% | — |
| LiveBench Language | 48.5% | — |
Frequently asked questions
Is DeepSeek-R1 better than Gemini 3.7 Flash?
Gemini 3.7 Flash is the stronger model overall, scoring 59.8 to 42.3 on the Noometry Index. DeepSeek-R1 costs 1.6× less per token, which makes it the better buy when Gemini 3.7 Flash's lead doesn't matter for your workload.
Which is cheaper, DeepSeek-R1 or Gemini 3.7 Flash?
DeepSeek-R1 is cheaper. It lists at $0.50 per million input tokens and $2.15 per million output tokens; Gemini 3.7 Flash lists at $0.75 and $3.75.
Is DeepSeek-R1 or Gemini 3.7 Flash better for coding?
Gemini 3.7 Flash scores higher on coding benchmarks: 56.2 versus 46.3 in the Noometry coding category.
Which has the bigger context window?
Gemini 3.7 Flash does, with 1.05M tokens against 164K.
How many benchmarks do DeepSeek-R1 and Gemini 3.7 Flash share?
26 benchmarks have published results for both models. DeepSeek-R1 has 52 scored results on Noometry and Gemini 3.7 Flash has 44.