Model comparison
DeepSeek-R1 vs GPT-4o mini
DeepSeek-R1 is the stronger model overall, scoring 42.3 to 25.5 on the Noometry Index. GPT-4o mini costs 3.5× less per token, which makes it the better buy when DeepSeek-R1's lead doesn't matter for your workload.
Last verified . 42 shared benchmarks.
DeepSeekSummary
- They share 42 benchmarks with published results for both. DeepSeek-R1 scores higher in 9 categories and GPT-4o mini in 0 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where DeepSeek-R1 leads 43.8 to 10.4.
- The biggest single-benchmark swing is Aider Polyglot: 71.4% for DeepSeek-R1 and 3.6% for GPT-4o mini.
- GPT-4o mini is cheaper at $0.15 / $0.60 per million input/output tokens, against $0.50 / $2.15 for DeepSeek-R1.
- DeepSeek-R1 accepts more context: 164K tokens versus 128K.
Side by side
| DeepSeek-R1 | GPT-4o mini | |
|---|---|---|
| Provider | DeepSeek | OpenAI |
| Noometry Index | 42.3 | 25.5 |
| Released | 2025-01-20 | 2024-07-18 |
| Weights | Proprietary | Proprietary |
| Context window | 164K | 128K |
| Max output | 64K | 16K |
| Input $ / M tokens | $0.50 | $0.15 |
| Output $ / M tokens | $2.15 | $0.60 |
| Results tracked | 52 | 60 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek-R1 leads
DeepSeek-R1: 46.3 (#68), GPT-4o mini: 22.0 (#335)
| Benchmark | DeepSeek-R1 | GPT-4o mini |
|---|---|---|
| Aider Polyglot | 71.4% | 3.6% |
| WeirdML | 41.6% | 11.8% |
| LiveBench Coding | 66.7% | 43.1% |
| LMArena Coding | 1427 | 1290 |
| SciCode | 35.7% | — |
| BigCodeBench Instruct | — | 46.1% |
| BigCodeBench Complete | — | 57.4% |
| ALE-Bench | 804.12 | — |
| AlgoTune | 1.7 | — |
| HumanEval+ | — | 83.5% |
| MBPP+ | — | 72.2% |
Agentic & Tool Use DeepSeek-R1 leads
DeepSeek-R1: 30.7 (#75), GPT-4o mini: 27.5 (#101)
| Benchmark | DeepSeek-R1 | GPT-4o mini |
|---|---|---|
| BALROG | 34.9% | 17.4% |
| DeepResearch Bench | 35.1% | — |
| METR Time Horizons | 53.8% | — |
Reasoning DeepSeek-R1 leads
DeepSeek-R1: 18.6 (#278), GPT-4o mini: 8.7 (#347)
| Benchmark | DeepSeek-R1 | GPT-4o mini |
|---|---|---|
| ARC-AGI-2 | 1.3% | 0% |
| SimpleBench | 40.8% | 10.7% |
| Kagi LLM Benchmark | 69.4% | 28.8% |
| LiveBench Reasoning | 83.2% | 32.8% |
| LMArena Hard Prompts | 1416 | 1267 |
| LiveBench Data Analysis | 69.8% | 50% |
| Epoch Capabilities Index | 141.29 | 126.56 |
| LiveBench | 71.6% | 41.3% |
| ARC-AGI-1 | 21.2% | — |
| CritPt | 1.1% | — |
| Chess Puzzles | — | 0% |
| Mystery Game Puzzles | — | 12% |
| DTBench | — | 54.4% |
| LMCA | — | 10.4% |
| ForecastBench | 60 | — |
| PIQA | — | 88.7% |
Math DeepSeek-R1 leads
DeepSeek-R1: 43.8 (#79), GPT-4o mini: 10.4 (#314)
| Benchmark | DeepSeek-R1 | GPT-4o mini |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 66.4% | 6.9% |
| Omni-MATH | 42.4% | 28% |
| LiveBench Math | 80.7% | 36.3% |
| LMArena Math | 1400 | 1267 |
| MATH Level 5 | 96.6% | 52.6% |
| FrontierMath (Tiers 1-3) | — | 0.7% |
| GSM8K | — | 91.3% |
Knowledge DeepSeek-R1 leads
DeepSeek-R1: 44.5 (#87), GPT-4o mini: 17.7 (#284)
| Benchmark | DeepSeek-R1 | GPT-4o mini |
|---|---|---|
| GPQA Diamond | 76.3% | 37.7% |
| MMLU-Pro | 79.3% | 60.3% |
| Confabulations | 12.7% | 37.2% |
| GPQA (HELM) | 66.6% | 36.8% |
| LMArena Expert | 1394 | 1235 |
| SimpleQA Verified | — | 8.3% |
| Vectara Hallucination Rate | 11.3% | — |
| BoolQ | — | 88.7% |
| MMLU | — | 81.8% |
Multimodal Not comparable
DeepSeek-R1: —, GPT-4o mini: 25.9 (#122)
| Benchmark | DeepSeek-R1 | GPT-4o mini |
|---|---|---|
| LMArena Vision | — | 1066 |
| Video-MME | — | 64.8% |
| GeoBench | — | 64% |
| VPCT | — | 34% |
Multilingual DeepSeek-R1 leads
DeepSeek-R1: 52.4 (#85), GPT-4o mini: 42.0 (#199)
| Benchmark | DeepSeek-R1 | GPT-4o mini |
|---|---|---|
| LMArena Non-English | 1412 | 1266 |
| LMArena Chinese | 1442 | 1265 |
| LMArena French | 1417 | 1297 |
| LMArena German | 1404 | 1272 |
| LMArena Japanese | 1391 | 1216 |
| LMArena Korean | 1360 | 1195 |
| LMArena Russian | 1423 | 1275 |
| LMArena Spanish | 1411 | 1276 |
Instruction Following DeepSeek-R1 leads
DeepSeek-R1: 72.0 (#143), GPT-4o mini: 61.9 (#239)
| Benchmark | DeepSeek-R1 | GPT-4o mini |
|---|---|---|
| LiveBench Instruction Following | 80.5% | 56.8% |
| IFEval | 78.4% | 78.2% |
| LMArena Instruction Following | 1382 | 1258 |
Long Context DeepSeek-R1 leads
DeepSeek-R1: 45.4 (#36), GPT-4o mini: 39.1 (#186)
| Benchmark | DeepSeek-R1 | GPT-4o mini |
|---|---|---|
| LMArena Longer Query | 1391 | 1289 |
| Fiction.LiveBench | 75% | — |
Writing & Preference DeepSeek-R1 leads
DeepSeek-R1: 61.4 (#88), GPT-4o mini: 39.5 (#248)
| Benchmark | DeepSeek-R1 | GPT-4o mini |
|---|---|---|
| LMArena Text | 1428 | 1286 |
| LMArena Creative Writing | 1405 | 1268 |
| Short-Story Creative Writing | 83% | 67.2% |
| EQ-Bench Creative Writing | 1500 | 873 |
| WildBench | 82.8% | 79.1% |
| LMArena Multi-Turn | 1405 | 1285 |
| LiveBench Language | 48.5% | 28.6% |
Frequently asked questions
Is DeepSeek-R1 better than GPT-4o mini?
DeepSeek-R1 is the stronger model overall, scoring 42.3 to 25.5 on the Noometry Index. GPT-4o mini costs 3.5× less per token, which makes it the better buy when DeepSeek-R1's lead doesn't matter for your workload.
Which is cheaper, DeepSeek-R1 or GPT-4o mini?
GPT-4o mini is cheaper. It lists at $0.15 per million input tokens and $0.60 per million output tokens; DeepSeek-R1 lists at $0.50 and $2.15.
Is DeepSeek-R1 or GPT-4o mini better for coding?
DeepSeek-R1 scores higher on coding benchmarks: 46.3 versus 22.0 in the Noometry coding category.
Which has the bigger context window?
DeepSeek-R1 does, with 164K tokens against 128K.
How many benchmarks do DeepSeek-R1 and GPT-4o mini share?
42 benchmarks have published results for both models. DeepSeek-R1 has 52 scored results on Noometry and GPT-4o mini has 60.