Model comparison
DeepSeek-R1 vs GPT-6 Luna
GPT-6 Luna is the stronger model overall, scoring 53.3 to 42.3 on the Noometry Index.
Last verified . 25 shared benchmarks.
DeepSeekSummary
- They share 25 benchmarks with published results for both. DeepSeek-R1 scores higher in 3 categories and GPT-6 Luna in 6 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-6 Luna leads 76.1 to 43.8.
- The biggest single-benchmark swing is ARC-AGI-1: 21.2% for DeepSeek-R1 and 86.7% for GPT-6 Luna.
- GPT-6 Luna is cheaper at $0.10 / $0.50 per million input/output tokens, against $0.50 / $2.15 for DeepSeek-R1.
- GPT-6 Luna accepts more context: 1.05M tokens versus 164K.
Side by side
| DeepSeek-R1 | GPT-6 Luna | |
|---|---|---|
| Provider | DeepSeek | OpenAI |
| Noometry Index | 42.3 | 53.3 |
| Released | 2025-01-20 | 2026-09-22 |
| Weights | Proprietary | Proprietary |
| Context window | 164K | 1.05M |
| Max output | 64K | 128K |
| Input $ / M tokens | $0.50 | $0.10 |
| Output $ / M tokens | $2.15 | $0.50 |
| Results tracked | 52 | 42 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-6 Luna leads
DeepSeek-R1: 46.3 (#68), GPT-6 Luna: 55.5 (#25)
| Benchmark | DeepSeek-R1 | GPT-6 Luna |
|---|---|---|
| SciCode | 35.7% | 54.6% |
| LMArena Coding | 1427 | 1439 |
| ALE-Bench | 804.12 | 1,577 |
| DeepSWE | — | 66.6% |
| FrontierCode | — | 42.4% |
| Aider Polyglot | 71.4% | — |
| LMArena WebDev | — | 1581 |
| WeirdML | 41.6% | — |
| LiveBench Coding | 66.7% | — |
| AlgoTune | 1.7 | — |
Agentic & Tool Use GPT-6 Luna leads
DeepSeek-R1: 30.7 (#75), GPT-6 Luna: 33.3 (#54)
| Benchmark | DeepSeek-R1 | GPT-6 Luna |
|---|---|---|
| APEX-Agents | — | 44.3% |
| DeepResearch Bench | 35.1% | — |
| BALROG | 34.9% | — |
| GDP.pdf | — | 23% |
| METR Time Horizons | 53.8% | — |
Reasoning GPT-6 Luna leads
DeepSeek-R1: 18.6 (#278), GPT-6 Luna: 48.2 (#41)
| Benchmark | DeepSeek-R1 | GPT-6 Luna |
|---|---|---|
| ARC-AGI-2 | 1.3% | 59.3% |
| ARC-AGI-1 | 21.2% | 86.7% |
| CritPt | 1.1% | 19.4% |
| LMArena Hard Prompts | 1416 | 1411 |
| Epoch Capabilities Index | 141.29 | 156.28 |
| SimpleBench | 40.8% | — |
| Kagi LLM Benchmark | 69.4% | — |
| NYT Connections (extended) | — | 68.7% |
| Chess Puzzles | — | 31% |
| LiveBench Reasoning | 83.2% | — |
| Mystery Game Puzzles | — | 7% |
| DTBench | — | 90.1% |
| LiveBench Data Analysis | 69.8% | — |
| LMCA | — | 44.5% |
| ForecastBench | 60 | — |
| LiveBench | 71.6% | — |
Math GPT-6 Luna leads
DeepSeek-R1: 43.8 (#79), GPT-6 Luna: 76.1 (#15)
| Benchmark | DeepSeek-R1 | GPT-6 Luna |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 66.4% | 98.9% |
| LMArena Math | 1400 | 1416 |
| FrontierMath (Tiers 1-3) | — | 78.9% |
| FrontierMath Tier 4 | — | 56.1% |
| ProofBench | — | 64% |
| Omni-MATH | 42.4% | — |
| LiveBench Math | 80.7% | — |
| MATH Level 5 | 96.6% | — |
Knowledge GPT-6 Luna leads
DeepSeek-R1: 44.5 (#87), GPT-6 Luna: 57.0 (#41)
| Benchmark | DeepSeek-R1 | GPT-6 Luna |
|---|---|---|
| GPQA Diamond | 76.3% | 90.5% |
| LMArena Expert | 1394 | 1444 |
| SimpleQA Verified | — | 41.4% |
| MMLU-Pro | 79.3% | — |
| Confabulations | 12.7% | — |
| Vectara Hallucination Rate | 11.3% | — |
| GPQA (HELM) | 66.6% | — |
Multimodal Not comparable
DeepSeek-R1: —, GPT-6 Luna: 42.4 (#30)
| Benchmark | DeepSeek-R1 | GPT-6 Luna |
|---|---|---|
| LMArena Vision | — | 1217 |
| Blueprint-Bench 2 | — | 31.2% |
| Furniture Assembly | — | 44.2% |
Multilingual DeepSeek-R1 leads
DeepSeek-R1: 52.4 (#85), GPT-6 Luna: 50.5 (#117)
| Benchmark | DeepSeek-R1 | GPT-6 Luna |
|---|---|---|
| LMArena Non-English | 1412 | 1386 |
| LMArena Chinese | 1442 | 1433 |
| LMArena French | 1417 | 1420 |
| LMArena German | 1404 | 1369 |
| LMArena Japanese | 1391 | 1369 |
| LMArena Korean | 1360 | 1360 |
| LMArena Russian | 1423 | 1394 |
| LMArena Spanish | 1411 | 1393 |
Instruction Following GPT-6 Luna leads
DeepSeek-R1: 72.0 (#143), GPT-6 Luna: 74.3 (#99)
| Benchmark | DeepSeek-R1 | GPT-6 Luna |
|---|---|---|
| LMArena Instruction Following | 1382 | 1409 |
| LiveBench Instruction Following | 80.5% | — |
| IFEval | 78.4% | — |
Long Context DeepSeek-R1 leads
DeepSeek-R1: 45.4 (#36), GPT-6 Luna: 43.0 (#111)
| Benchmark | DeepSeek-R1 | GPT-6 Luna |
|---|---|---|
| LMArena Longer Query | 1391 | 1409 |
| Fiction.LiveBench | 75% | — |
Writing & Preference DeepSeek-R1 leads
DeepSeek-R1: 61.4 (#88), GPT-6 Luna: 58.3 (#119)
| Benchmark | DeepSeek-R1 | GPT-6 Luna |
|---|---|---|
| LMArena Text | 1428 | 1391 |
| LMArena Creative Writing | 1405 | 1363 |
| LMArena Multi-Turn | 1405 | 1396 |
| Short-Story Creative Writing | 83% | — |
| EQ-Bench Creative Writing | 1500 | — |
| WildBench | 82.8% | — |
| LiveBench Language | 48.5% | — |
Frequently asked questions
Is DeepSeek-R1 better than GPT-6 Luna?
GPT-6 Luna is the stronger model overall, scoring 53.3 to 42.3 on the Noometry Index.
Which is cheaper, DeepSeek-R1 or GPT-6 Luna?
GPT-6 Luna is cheaper. It lists at $0.10 per million input tokens and $0.50 per million output tokens; DeepSeek-R1 lists at $0.50 and $2.15.
Is DeepSeek-R1 or GPT-6 Luna better for coding?
GPT-6 Luna scores higher on coding benchmarks: 55.5 versus 46.3 in the Noometry coding category.
Which has the bigger context window?
GPT-6 Luna does, with 1.05M tokens against 164K.
How many benchmarks do DeepSeek-R1 and GPT-6 Luna share?
25 benchmarks have published results for both models. DeepSeek-R1 has 52 scored results on Noometry and GPT-6 Luna has 42.