Model comparison
DeepSeek V4 Flash vs GPT-5.6 Luna
DeepSeek V4 Flash and GPT-5.6 Luna score almost the same on the Noometry Index (53.6 vs 54.6), so choose on price, context window or the category you care about most.
Last verified . 40 shared benchmarks.
DeepSeekSummary
- They share 40 benchmarks with published results for both. DeepSeek V4 Flash scores higher in 2 categories and GPT-5.6 Luna in 6 categories; 5 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-5.6 Luna leads 77.7 to 60.3.
- The biggest single-benchmark swing is FrontierMath Tier 4: 24.4% for DeepSeek V4 Flash and 61% for GPT-5.6 Luna.
- DeepSeek V4 Flash is cheaper at $0.15 / $0.60 per million input/output tokens, against $0.20 / $1.20 for GPT-5.6 Luna.
- GPT-5.6 Luna accepts more context: 1.05M tokens versus 1M.
- DeepSeek V4 Flash has downloadable open weights; the other is API-only.
Side by side
| DeepSeek V4 Flash | GPT-5.6 Luna | |
|---|---|---|
| Provider | DeepSeek | OpenAI |
| Noometry Index | 53.6 | 54.6 |
| Released | 2026-04-24 | 2026-07-09 |
| Weights | Open | Proprietary |
| Context window | 1M | 1.05M |
| Max output | 393K | 128K |
| Input $ / M tokens | $0.15 | $0.20 |
| Output $ / M tokens | $0.60 | $1.20 |
| Results tracked | 41 | 52 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5.6 Luna leads
DeepSeek V4 Flash: 47.9 (#59), GPT-5.6 Luna: 54.5 (#28)
| Benchmark | DeepSeek V4 Flash | GPT-5.6 Luna |
|---|---|---|
| FrontierCode | 18.8% | 39.8% |
| LMArena WebDev | 1582 | 1519 |
| SciCode | 49.9% | 53.6% |
| WeirdML | 63% | 60.9% |
| LMArena Coding | 1457 | 1466 |
| ALE-Bench | 1,306 | 1,667 |
| DeepSWE | — | 67.2% |
| CursorBench | — | 35.9% |
Agentic & Tool Use Not comparable
DeepSeek V4 Flash: —, GPT-5.6 Luna: 34.4 (#45)
| Benchmark | DeepSeek V4 Flash | GPT-5.6 Luna |
|---|---|---|
| APEX-Agents | — | 43% |
| BALROG | — | 45.6% |
| GDP.pdf | — | 22.7% |
| Vending-Bench 2 | — | 4,095 |
Reasoning DeepSeek V4 Flash leads
DeepSeek V4 Flash: 53.7 (#30), GPT-5.6 Luna: 47.6 (#43)
| Benchmark | DeepSeek V4 Flash | GPT-5.6 Luna |
|---|---|---|
| ARC-AGI-2 | 61.4% | 59.5% |
| SimpleBench | 61.1% | 46.8% |
| Kagi LLM Benchmark | 52.2% | 49.1% |
| NYT Connections (extended) | 89.6% | 69.4% |
| ARC-AGI-1 | 89% | 88% |
| CritPt | 16.6% | 20.6% |
| Chess Puzzles | 33% | 40% |
| LMArena Hard Prompts | 1444 | 1451 |
| Mystery Game Puzzles | 34% | 21% |
| DTBench | 90.9% | 89.1% |
| LMCA | 41.7% | 48.5% |
| Epoch Capabilities Index | 154.49 | 156.39 |
| Surface Evolver Bench | — | 61.9% |
Math GPT-5.6 Luna leads
DeepSeek V4 Flash: 60.3 (#37), GPT-5.6 Luna: 77.7 (#14)
| Benchmark | DeepSeek V4 Flash | GPT-5.6 Luna |
|---|---|---|
| FrontierMath (Tiers 1-3) | 57.5% | 82.1% |
| FrontierMath Tier 4 | 24.4% | 61% |
| OTIS Mock AIME 2024-2025 | 94.4% | 98.3% |
| ProofBench | 56% | 60% |
| LMArena Math | 1427 | 1458 |
| MathArena Final-Answer Competitions | 76.5% | — |
Knowledge GPT-5.6 Luna leads
DeepSeek V4 Flash: 55.4 (#48), GPT-5.6 Luna: 58.5 (#34)
| Benchmark | DeepSeek V4 Flash | GPT-5.6 Luna |
|---|---|---|
| GPQA Diamond | 91% | 91.6% |
| SimpleQA Verified | 33.6% | 41% |
| LMArena Expert | 1441 | 1478 |
Multimodal Not comparable
DeepSeek V4 Flash: —, GPT-5.6 Luna: 42.7 (#28)
| Benchmark | DeepSeek V4 Flash | GPT-5.6 Luna |
|---|---|---|
| LMArena Vision | — | 1258 |
| Blueprint-Bench 2 | — | 22.6% |
| Furniture Assembly | — | 42.5% |
| LMArena Document | — | 1457 |
Multilingual Too close to call
DeepSeek V4 Flash: 53.0 (#72), GPT-5.6 Luna: 52.8 (#78)
| Benchmark | DeepSeek V4 Flash | GPT-5.6 Luna |
|---|---|---|
| LMArena Non-English | 1420 | 1417 |
| LMArena Chinese | 1468 | 1470 |
| LMArena French | 1439 | 1456 |
| LMArena German | 1418 | 1454 |
| LMArena Japanese | 1406 | 1411 |
| LMArena Korean | 1384 | 1415 |
| LMArena Russian | 1428 | 1428 |
| LMArena Spanish | 1436 | 1448 |
Instruction Following Too close to call
DeepSeek V4 Flash: 74.9 (#81), GPT-5.6 Luna: 75.6 (#57)
| Benchmark | DeepSeek V4 Flash | GPT-5.6 Luna |
|---|---|---|
| LMArena Instruction Following | 1421 | 1437 |
Long Context Too close to call
DeepSeek V4 Flash: 43.8 (#85), GPT-5.6 Luna: 43.9 (#82)
| Benchmark | DeepSeek V4 Flash | GPT-5.6 Luna |
|---|---|---|
| LMArena Longer Query | 1434 | 1436 |
Writing & Preference GPT-5.6 Luna leads
DeepSeek V4 Flash: 63.8 (#61), GPT-5.6 Luna: 68.0 (#29)
| Benchmark | DeepSeek V4 Flash | GPT-5.6 Luna |
|---|---|---|
| LMArena Text | 1432 | 1431 |
| LMArena Creative Writing | 1403 | 1396 |
| EQ-Bench Creative Writing | 1559 | 1829 |
| LMArena Multi-Turn | 1449 | 1434 |
| EQ-Bench 4 | — | 1156 |
Frequently asked questions
Is DeepSeek V4 Flash better than GPT-5.6 Luna?
DeepSeek V4 Flash and GPT-5.6 Luna score almost the same on the Noometry Index (53.6 vs 54.6), so choose on price, context window or the category you care about most.
Which is cheaper, DeepSeek V4 Flash or GPT-5.6 Luna?
DeepSeek V4 Flash is cheaper. It lists at $0.15 per million input tokens and $0.60 per million output tokens; GPT-5.6 Luna lists at $0.20 and $1.20.
Is DeepSeek V4 Flash or GPT-5.6 Luna better for coding?
GPT-5.6 Luna scores higher on coding benchmarks: 54.5 versus 47.9 in the Noometry coding category.
Which has the bigger context window?
GPT-5.6 Luna does, with 1.05M tokens against 1M.
How many benchmarks do DeepSeek V4 Flash and GPT-5.6 Luna share?
40 benchmarks have published results for both models. DeepSeek V4 Flash has 41 scored results on Noometry and GPT-5.6 Luna has 52.