Model comparison
DeepSeek V4 Pro vs o1
DeepSeek V4 Pro is the stronger model overall, scoring 54.3 to 40.9 on the Noometry Index.
Last verified . 27 shared benchmarks.
DeepSeekSummary
- They share 27 benchmarks with published results for both. DeepSeek V4 Pro scores higher in 8 categories and o1 in 1 category; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where DeepSeek V4 Pro leads 64.8 to 36.1.
- The biggest single-benchmark swing is ARC-AGI-1: 90.5% for DeepSeek V4 Pro and 30.7% for o1.
- DeepSeek V4 Pro is cheaper at $0.66 / $1.98 per million input/output tokens, against $15 / $60 for o1.
- DeepSeek V4 Pro accepts more context: 1M tokens versus 200K.
- DeepSeek V4 Pro has downloadable open weights; the other is API-only.
Side by side
| DeepSeek V4 Pro | o1 | |
|---|---|---|
| Provider | DeepSeek | OpenAI |
| Noometry Index | 54.3 | 40.9 |
| Released | 2026-04-24 | 2024-09-12 |
| Weights | Open | Proprietary |
| Context window | 1M | 200K |
| Max output | 393K | 100K |
| Input $ / M tokens | $0.66 | $15 |
| Output $ / M tokens | $1.98 | $60 |
| Results tracked | 48 | 52 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek V4 Pro leads
DeepSeek V4 Pro: 52.4 (#34), o1: 46.1 (#70)
| Benchmark | DeepSeek V4 Pro | o1 |
|---|---|---|
| WeirdML | 66.2% | 47.6% |
| LMArena Coding | 1470 | 1367 |
| SWE-bench Verified | 77.6% | — |
| FrontierCode | 28.6% | — |
| Aider Polyglot | — | 61.7% |
| LMArena WebDev | 1582 | — |
| SciCode | 51% | — |
| LiveBench Coding | — | 69.7% |
| CadEval | — | 56% |
| ALE-Bench | 1,403 | — |
| HumanEval+ | — | 89% |
| MBPP+ | — | 80.2% |
Agentic & Tool Use DeepSeek V4 Pro leads
DeepSeek V4 Pro: 32.8 (#58), o1: 24.6 (#117)
| Benchmark | DeepSeek V4 Pro | o1 |
|---|---|---|
| APEX-Agents | 47.3% | — |
| Cybench | — | 10% |
| METR Time Horizons | — | 51.1% |
| Vending-Bench 2 | 3,285 | — |
Reasoning DeepSeek V4 Pro leads
DeepSeek V4 Pro: 56.5 (#24), o1: 27.9 (#111)
| Benchmark | DeepSeek V4 Pro | o1 |
|---|---|---|
| ARC-AGI-1 | 90.5% | 30.7% |
| Chess Puzzles | 47% | 15% |
| LMArena Hard Prompts | 1461 | 1371 |
| DTBench | 93.9% | 74.7% |
| LMCA | 45.5% | 22.3% |
| Epoch Capabilities Index | 155.31 | 141.91 |
| ARC-AGI-2 | 61.3% | — |
| SimpleBench | — | 41.7% |
| Kagi LLM Benchmark | 53.5% | — |
| NYT Connections (extended) | 91.3% | — |
| CritPt | 18% | — |
| EnigmaEval | — | 5.7% |
| LiveBench Reasoning | — | 91.6% |
| Mystery Game Puzzles | 43% | — |
| LiveBench Data Analysis | — | 65.5% |
| Surface Evolver Bench | 40% | — |
| ForecastBench | 56.1 | — |
| LiveBench | — | 75.7% |
Math DeepSeek V4 Pro leads
DeepSeek V4 Pro: 64.8 (#30), o1: 36.1 (#175)
| Benchmark | DeepSeek V4 Pro | o1 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 64.6% | 14.7% |
| OTIS Mock AIME 2024-2025 | 98.6% | 73.3% |
| LMArena Math | 1455 | 1388 |
| FrontierMath Tier 4 | 26.8% | — |
| MathArena Final-Answer Competitions | 76.6% | — |
| ProofBench | 50% | — |
| LiveBench Math | — | 80.3% |
| MATH Level 5 | — | 94.7% |
| FrontierMath (Feb 2025 set) | — | 9.3% |
Knowledge DeepSeek V4 Pro leads
DeepSeek V4 Pro: 59.5 (#31), o1: 41.5 (#110)
| Benchmark | DeepSeek V4 Pro | o1 |
|---|---|---|
| GPQA Diamond | 91.7% | 76.8% |
| SimpleQA Verified | 52.9% | 41.1% |
| LMArena Expert | 1464 | 1361 |
| Humanity's Last Exam | — | 8% |
| Confabulations | — | 11.7% |
| Vectara Hallucination Rate | 8.6% | — |
Multimodal Not comparable
DeepSeek V4 Pro: —, o1: 34.2 (#93)
| Benchmark | DeepSeek V4 Pro | o1 |
|---|---|---|
| LMArena Vision | — | 1168 |
| GeoBench | — | 80% |
| VPCT | — | 37% |
| SpatialViz-Bench | — | 41.4% |
Multilingual DeepSeek V4 Pro leads
DeepSeek V4 Pro: 54.4 (#45), o1: 48.6 (#142)
| Benchmark | DeepSeek V4 Pro | o1 |
|---|---|---|
| LMArena Non-English | 1439 | 1358 |
| LMArena Chinese | 1486 | 1394 |
| LMArena French | 1472 | 1344 |
| LMArena German | 1458 | 1337 |
| LMArena Japanese | 1445 | 1346 |
| LMArena Korean | 1447 | 1396 |
| LMArena Russian | 1453 | 1356 |
| LMArena Spanish | 1458 | 1345 |
Instruction Following DeepSeek V4 Pro leads
DeepSeek V4 Pro: 76.1 (#47), o1: 74.8 (#86)
| Benchmark | DeepSeek V4 Pro | o1 |
|---|---|---|
| LMArena Instruction Following | 1448 | 1367 |
| LiveBench Instruction Following | — | 81.5% |
Long Context o1 leads
DeepSeek V4 Pro: 45.0 (#51), o1: 50.3 (#9)
| Benchmark | DeepSeek V4 Pro | o1 |
|---|---|---|
| LMArena Longer Query | 1458 | 1378 |
| Fiction.LiveBench | — | 83.3% |
| CL-bench Life | 13.5% | — |
Writing & Preference DeepSeek V4 Pro leads
DeepSeek V4 Pro: 65.5 (#46), o1: 55.6 (#144)
| Benchmark | DeepSeek V4 Pro | o1 |
|---|---|---|
| LMArena Text | 1451 | 1366 |
| LMArena Creative Writing | 1446 | 1348 |
| LMArena Multi-Turn | 1467 | 1369 |
| Short-Story Creative Writing | — | 70.2% |
| EQ-Bench Creative Writing | 1553 | — |
| EQ-Bench 4 | 1166 | — |
| LiveBench Language | — | 65.4% |
Frequently asked questions
Is DeepSeek V4 Pro better than o1?
DeepSeek V4 Pro is the stronger model overall, scoring 54.3 to 40.9 on the Noometry Index.
Which is cheaper, DeepSeek V4 Pro or o1?
DeepSeek V4 Pro is cheaper. It lists at $0.66 per million input tokens and $1.98 per million output tokens; o1 lists at $15 and $60.
Is DeepSeek V4 Pro or o1 better for coding?
DeepSeek V4 Pro scores higher on coding benchmarks: 52.4 versus 46.1 in the Noometry coding category.
Which has the bigger context window?
DeepSeek V4 Pro does, with 1M tokens against 200K.
How many benchmarks do DeepSeek V4 Pro and o1 share?
27 benchmarks have published results for both models. DeepSeek V4 Pro has 48 scored results on Noometry and o1 has 52.