Model comparison
DeepSeek V4 Flash vs GPT-4.5
DeepSeek V4 Flash is the stronger model overall, scoring 53.6 to 37.2 on the Noometry Index.
Last verified . 24 shared benchmarks.
DeepSeekSummary
- They share 24 benchmarks with published results for both. DeepSeek V4 Flash scores higher in 8 categories and GPT-4.5 in 0 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where DeepSeek V4 Flash leads 53.7 to 13.9.
- The biggest single-benchmark swing is ARC-AGI-1: 89% for DeepSeek V4 Flash and 10.3% for GPT-4.5.
- DeepSeek V4 Flash has downloadable open weights; the other is API-only.
Side by side
| DeepSeek V4 Flash | GPT-4.5 | |
|---|---|---|
| Provider | DeepSeek | OpenAI |
| Noometry Index | 53.6 | 37.2 |
| Released | 2026-04-24 | 2025-02-27 |
| Weights | Open | Proprietary |
| Context window | 1M | — |
| Max output | 393K | — |
| Input $ / M tokens | $0.15 | — |
| Output $ / M tokens | $0.60 | — |
| Results tracked | 41 | 42 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek V4 Flash leads
DeepSeek V4 Flash: 47.9 (#59), GPT-4.5: 42.2 (#109)
| Benchmark | DeepSeek V4 Flash | GPT-4.5 |
|---|---|---|
| WeirdML | 63% | 39.4% |
| LMArena Coding | 1457 | 1396 |
| FrontierCode | 18.8% | — |
| Aider Polyglot | — | 44.9% |
| LMArena WebDev | 1582 | — |
| SciCode | 49.9% | — |
| LiveBench Coding | — | 75.2% |
| ALE-Bench | 1,306 | — |
Agentic & Tool Use Not comparable
DeepSeek V4 Flash: —, GPT-4.5: 27.9 (#97)
| Benchmark | DeepSeek V4 Flash | GPT-4.5 |
|---|---|---|
| Cybench | — | 17.5% |
Reasoning DeepSeek V4 Flash leads
DeepSeek V4 Flash: 53.7 (#30), GPT-4.5: 13.9 (#330)
| Benchmark | DeepSeek V4 Flash | GPT-4.5 |
|---|---|---|
| ARC-AGI-2 | 61.4% | 0.8% |
| SimpleBench | 61.1% | 34.5% |
| ARC-AGI-1 | 89% | 10.3% |
| LMArena Hard Prompts | 1444 | 1403 |
| Epoch Capabilities Index | 154.49 | 136.74 |
| Kagi LLM Benchmark | 52.2% | — |
| NYT Connections (extended) | 89.6% | — |
| CritPt | 16.6% | — |
| Chess Puzzles | 33% | — |
| EnigmaEval | — | 3.2% |
| LiveBench Reasoning | — | 71.1% |
| Mystery Game Puzzles | 34% | — |
| DTBench | 90.9% | — |
| LiveBench Data Analysis | — | 64.3% |
| LMCA | 41.7% | — |
| ForecastBench | — | 61.7 |
| LiveBench | — | 69% |
Math DeepSeek V4 Flash leads
DeepSeek V4 Flash: 60.3 (#37), GPT-4.5: 32.6 (#211)
| Benchmark | DeepSeek V4 Flash | GPT-4.5 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 94.4% | 37.8% |
| LMArena Math | 1427 | 1412 |
| FrontierMath (Tiers 1-3) | 57.5% | — |
| FrontierMath Tier 4 | 24.4% | — |
| MathArena Final-Answer Competitions | 76.5% | — |
| ProofBench | 56% | — |
| LiveBench Math | — | 69.3% |
| MATH Level 5 | — | 78.6% |
Knowledge DeepSeek V4 Flash leads
DeepSeek V4 Flash: 55.4 (#48), GPT-4.5: 32.5 (#211)
| Benchmark | DeepSeek V4 Flash | GPT-4.5 |
|---|---|---|
| GPQA Diamond | 91% | 68.7% |
| LMArena Expert | 1441 | 1394 |
| Humanity's Last Exam | — | 5.4% |
| SimpleQA Verified | 33.6% | — |
| Confabulations | — | 13.6% |
Multimodal Not comparable
DeepSeek V4 Flash: —, GPT-4.5: 37.6 (#71)
| Benchmark | DeepSeek V4 Flash | GPT-4.5 |
|---|---|---|
| LMArena Vision | — | 1195 |
| VPCT | — | 45% |
Multilingual Too close to call
DeepSeek V4 Flash: 53.0 (#72), GPT-4.5: 52.5 (#83)
| Benchmark | DeepSeek V4 Flash | GPT-4.5 |
|---|---|---|
| LMArena Non-English | 1420 | 1413 |
| LMArena Chinese | 1468 | 1421 |
| LMArena French | 1439 | 1418 |
| LMArena German | 1418 | 1457 |
| LMArena Japanese | 1406 | 1416 |
| LMArena Korean | 1384 | 1392 |
| LMArena Russian | 1428 | 1419 |
| LMArena Spanish | 1436 | — |
Instruction Following DeepSeek V4 Flash leads
DeepSeek V4 Flash: 74.9 (#81), GPT-4.5: 72.6 (#134)
| Benchmark | DeepSeek V4 Flash | GPT-4.5 |
|---|---|---|
| LMArena Instruction Following | 1421 | 1404 |
| LiveBench Instruction Following | — | 72.3% |
Long Context DeepSeek V4 Flash leads
DeepSeek V4 Flash: 43.8 (#85), GPT-4.5: 40.4 (#155)
| Benchmark | DeepSeek V4 Flash | GPT-4.5 |
|---|---|---|
| LMArena Longer Query | 1434 | 1406 |
| Fiction.LiveBench | — | 63.9% |
Writing & Preference DeepSeek V4 Flash leads
DeepSeek V4 Flash: 63.8 (#61), GPT-4.5: 56.9 (#134)
| Benchmark | DeepSeek V4 Flash | GPT-4.5 |
|---|---|---|
| LMArena Text | 1432 | 1417 |
| LMArena Creative Writing | 1403 | 1394 |
| EQ-Bench Creative Writing | 1559 | 1258 |
| LMArena Multi-Turn | 1449 | 1444 |
| Short-Story Creative Writing | — | 75.6% |
| LiveBench Language | — | 61.5% |
Frequently asked questions
Is DeepSeek V4 Flash better than GPT-4.5?
DeepSeek V4 Flash is the stronger model overall, scoring 53.6 to 37.2 on the Noometry Index.
Is DeepSeek V4 Flash or GPT-4.5 better for coding?
DeepSeek V4 Flash scores higher on coding benchmarks: 47.9 versus 42.2 in the Noometry coding category.
How many benchmarks do DeepSeek V4 Flash and GPT-4.5 share?
24 benchmarks have published results for both models. DeepSeek V4 Flash has 41 scored results on Noometry and GPT-4.5 has 42.