Model comparison
DeepSeek-V3 vs GPT-6 Luna
GPT-6 Luna is the stronger model overall, scoring 53.3 to 39.5 on the Noometry Index.
Last verified . 24 shared benchmarks.
DeepSeekSummary
- They share 24 benchmarks with published results for both. DeepSeek-V3 scores higher in 0 categories and GPT-6 Luna in 8 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-6 Luna leads 76.1 to 32.1.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 37.8% for DeepSeek-V3 and 98.9% for GPT-6 Luna.
- GPT-6 Luna is cheaper at $0.10 / $0.50 per million input/output tokens, against $0.24 / $0.90 for DeepSeek-V3.
- GPT-6 Luna accepts more context: 1.05M tokens versus 164K.
- DeepSeek-V3 has downloadable open weights; the other is API-only.
Side by side
| DeepSeek-V3 | GPT-6 Luna | |
|---|---|---|
| Provider | DeepSeek | OpenAI |
| Noometry Index | 39.5 | 53.3 |
| Released | 2024-12-26 | 2026-09-22 |
| Weights | Open | Proprietary |
| Context window | 164K | 1.05M |
| Max output | 164K | 128K |
| Input $ / M tokens | $0.24 | $0.10 |
| Output $ / M tokens | $0.90 | $0.50 |
| Results tracked | 60 | 42 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-6 Luna leads
DeepSeek-V3: 42.3 (#106), GPT-6 Luna: 55.5 (#25)
| Benchmark | DeepSeek-V3 | GPT-6 Luna |
|---|---|---|
| SciCode | 35.8% | 54.6% |
| LMArena Coding | 1368 | 1439 |
| DeepSWE | — | 66.6% |
| FrontierCode | — | 42.4% |
| Aider Polyglot | 55.1% | — |
| LMArena WebDev | — | 1581 |
| WeirdML | 36.1% | — |
| BigCodeBench Instruct | 50% | — |
| LiveBench Coding | 70.9% | — |
| BigCodeBench Complete | 62.2% | — |
| ALE-Bench | — | 1,577 |
| HumanEval+ | 86.6% | — |
| MBPP+ | 73% | — |
Agentic & Tool Use Not comparable
DeepSeek-V3: —, GPT-6 Luna: 33.3 (#54)
| Benchmark | DeepSeek-V3 | GPT-6 Luna |
|---|---|---|
| APEX-Agents | — | 44.3% |
| GDP.pdf | — | 23% |
| METR Time Horizons | 49.6% | — |
Reasoning GPT-6 Luna leads
DeepSeek-V3: 20.5 (#236), GPT-6 Luna: 48.2 (#41)
| Benchmark | DeepSeek-V3 | GPT-6 Luna |
|---|---|---|
| CritPt | 0% | 19.4% |
| LMArena Hard Prompts | 1365 | 1411 |
| DTBench | 64.8% | 90.1% |
| LMCA | 15.5% | 44.5% |
| Epoch Capabilities Index | 135.94 | 156.28 |
| ARC-AGI-2 | — | 59.3% |
| SimpleBench | 27.2% | — |
| Kagi LLM Benchmark | 52.3% | — |
| NYT Connections (extended) | — | 68.7% |
| ARC-AGI-1 | — | 86.7% |
| Chess Puzzles | — | 31% |
| LiveBench Reasoning | 65.8% | — |
| Mystery Game Puzzles | — | 7% |
| LiveBench Data Analysis | 60.9% | — |
| BIG-Bench Hard | 87.5% | — |
| ForecastBench | 59.1 | — |
| HellaSwag | 88.9% | — |
| LiveBench | 66.9% | — |
| PIQA | 84.7% | — |
| WinoGrande | 85.2% | — |
Math GPT-6 Luna leads
DeepSeek-V3: 32.1 (#219), GPT-6 Luna: 76.1 (#15)
| Benchmark | DeepSeek-V3 | GPT-6 Luna |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 37.8% | 98.9% |
| LMArena Math | 1373 | 1416 |
| FrontierMath (Tiers 1-3) | — | 78.9% |
| FrontierMath Tier 4 | — | 56.1% |
| ProofBench | — | 64% |
| Omni-MATH | 40.3% | — |
| LiveBench Math | 73.5% | — |
| MATH Level 5 | 75.5% | — |
| FrontierMath (Feb 2025 set) | 1.7% | — |
Knowledge GPT-6 Luna leads
DeepSeek-V3: 37.5 (#155), GPT-6 Luna: 57.0 (#41)
| Benchmark | DeepSeek-V3 | GPT-6 Luna |
|---|---|---|
| GPQA Diamond | 67.6% | 90.5% |
| LMArena Expert | 1351 | 1444 |
| SimpleQA Verified | — | 41.4% |
| MMLU-Pro | 72.3% | — |
| Confabulations | 26.1% | — |
| Vectara Hallucination Rate | 6.1% | — |
| GPQA (HELM) | 53.8% | — |
| ARC (AI2) Challenge | 95.3% | — |
| MMLU | 87.2% | — |
| TriviaQA | 82.9% | — |
Multimodal Not comparable
DeepSeek-V3: —, GPT-6 Luna: 42.4 (#30)
| Benchmark | DeepSeek-V3 | GPT-6 Luna |
|---|---|---|
| LMArena Vision | — | 1217 |
| Blueprint-Bench 2 | — | 31.2% |
| Furniture Assembly | — | 44.2% |
Multilingual GPT-6 Luna leads
DeepSeek-V3: 48.5 (#143), GPT-6 Luna: 50.5 (#117)
| Benchmark | DeepSeek-V3 | GPT-6 Luna |
|---|---|---|
| LMArena Non-English | 1358 | 1386 |
| LMArena Chinese | 1391 | 1433 |
| LMArena French | 1385 | 1420 |
| LMArena German | 1374 | 1369 |
| LMArena Japanese | 1333 | 1369 |
| LMArena Korean | 1319 | 1360 |
| LMArena Russian | 1373 | 1394 |
| LMArena Spanish | 1358 | 1393 |
Instruction Following GPT-6 Luna leads
DeepSeek-V3: 72.8 (#130), GPT-6 Luna: 74.3 (#99)
| Benchmark | DeepSeek-V3 | GPT-6 Luna |
|---|---|---|
| LMArena Instruction Following | 1345 | 1409 |
| LiveBench Instruction Following | 81.5% | — |
| IFEval | 83.2% | — |
Long Context GPT-6 Luna leads
DeepSeek-V3: 34.0 (#253), GPT-6 Luna: 43.0 (#111)
| Benchmark | DeepSeek-V3 | GPT-6 Luna |
|---|---|---|
| LMArena Longer Query | 1352 | 1409 |
| Fiction.LiveBench | 50% | — |
Writing & Preference Too close to call
DeepSeek-V3: 57.4 (#130), GPT-6 Luna: 58.3 (#119)
| Benchmark | DeepSeek-V3 | GPT-6 Luna |
|---|---|---|
| LMArena Text | 1375 | 1391 |
| LMArena Creative Writing | 1364 | 1363 |
| LMArena Multi-Turn | 1389 | 1396 |
| Short-Story Creative Writing | 77% | — |
| EQ-Bench Creative Writing | 1472 | — |
| WildBench | 83% | — |
| LiveBench Language | 49.1% | — |
Frequently asked questions
Is DeepSeek-V3 better than GPT-6 Luna?
GPT-6 Luna is the stronger model overall, scoring 53.3 to 39.5 on the Noometry Index.
Which is cheaper, DeepSeek-V3 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-V3 lists at $0.24 and $0.90.
Is DeepSeek-V3 or GPT-6 Luna better for coding?
GPT-6 Luna scores higher on coding benchmarks: 55.5 versus 42.3 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-V3 and GPT-6 Luna share?
24 benchmarks have published results for both models. DeepSeek-V3 has 60 scored results on Noometry and GPT-6 Luna has 42.