Model comparison
DeepSeek V4 Pro vs o3-pro
DeepSeek V4 Pro is the stronger model overall, scoring 54.3 to 42.9 on the Noometry Index.
Last verified . 8 shared benchmarks.
DeepSeekSummary
- They share 8 benchmarks with published results for both. DeepSeek V4 Pro scores higher in 3 categories and o3-pro in 2 categories; 5 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where DeepSeek V4 Pro leads 56.5 to 23.8.
- The biggest single-benchmark swing is ARC-AGI-2: 61.3% for DeepSeek V4 Pro and 4.9% for o3-pro.
- DeepSeek V4 Pro is cheaper at $0.66 / $1.98 per million input/output tokens, against $20 / $80 for o3-pro.
- 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 | o3-pro | |
|---|---|---|
| Provider | DeepSeek | OpenAI |
| Noometry Index | 54.3 | 42.9 |
| Released | 2026-04-24 | 2025-06-10 |
| Weights | Open | Proprietary |
| Context window | 1M | 200K |
| Max output | 393K | 100K |
| Input $ / M tokens | $0.66 | $20 |
| Output $ / M tokens | $1.98 | $80 |
| Results tracked | 48 | 12 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding o3-pro leads
DeepSeek V4 Pro: 52.4 (#34), o3-pro: 55.5 (#24)
| Benchmark | DeepSeek V4 Pro | o3-pro |
|---|---|---|
| WeirdML | 66.2% | 58.2% |
| SWE-bench Verified | 77.6% | — |
| FrontierCode | 28.6% | — |
| Aider Polyglot | — | 84.9% |
| LMArena WebDev | 1582 | — |
| SciCode | 51% | — |
| LMArena Coding | 1470 | — |
| ALE-Bench | 1,403 | — |
Agentic & Tool Use Not comparable
DeepSeek V4 Pro: 32.8 (#58), o3-pro: —
| Benchmark | DeepSeek V4 Pro | o3-pro |
|---|---|---|
| APEX-Agents | 47.3% | — |
| Vending-Bench 2 | 3,285 | — |
Reasoning DeepSeek V4 Pro leads
DeepSeek V4 Pro: 56.5 (#24), o3-pro: 23.8 (#171)
| Benchmark | DeepSeek V4 Pro | o3-pro |
|---|---|---|
| ARC-AGI-2 | 61.3% | 4.9% |
| Kagi LLM Benchmark | 53.5% | 72.1% |
| ARC-AGI-1 | 90.5% | 59.3% |
| DTBench | 93.9% | 86.9% |
| LMCA | 45.5% | 38.5% |
| Epoch Capabilities Index | 155.31 | 147.42 |
| NYT Connections (extended) | 91.3% | — |
| CritPt | 18% | — |
| Chess Puzzles | 47% | — |
| LMArena Hard Prompts | 1461 | — |
| Mystery Game Puzzles | 43% | — |
| Surface Evolver Bench | 40% | — |
| ForecastBench | 56.1 | — |
Math Not comparable
DeepSeek V4 Pro: 64.8 (#30), o3-pro: —
| Benchmark | DeepSeek V4 Pro | o3-pro |
|---|---|---|
| FrontierMath (Tiers 1-3) | 64.6% | — |
| FrontierMath Tier 4 | 26.8% | — |
| MathArena Final-Answer Competitions | 76.6% | — |
| OTIS Mock AIME 2024-2025 | 98.6% | — |
| ProofBench | 50% | — |
| LMArena Math | 1455 | — |
Knowledge DeepSeek V4 Pro leads
DeepSeek V4 Pro: 59.5 (#31), o3-pro: 29.5 (#238)
| Benchmark | DeepSeek V4 Pro | o3-pro |
|---|---|---|
| Vectara Hallucination Rate | 8.6% | 23.3% |
| GPQA Diamond | 91.7% | — |
| SimpleQA Verified | 52.9% | — |
| Confabulations | — | 14.2% |
| LMArena Expert | 1464 | — |
Multilingual Not comparable
DeepSeek V4 Pro: 54.4 (#45), o3-pro: —
| Benchmark | DeepSeek V4 Pro | o3-pro |
|---|---|---|
| LMArena Non-English | 1439 | — |
| LMArena Chinese | 1486 | — |
| LMArena French | 1472 | — |
| LMArena German | 1458 | — |
| LMArena Japanese | 1445 | — |
| LMArena Korean | 1447 | — |
| LMArena Russian | 1453 | — |
| LMArena Spanish | 1458 | — |
Instruction Following Not comparable
DeepSeek V4 Pro: 76.1 (#47), o3-pro: —
| Benchmark | DeepSeek V4 Pro | o3-pro |
|---|---|---|
| LMArena Instruction Following | 1448 | — |
Long Context o3-pro leads
DeepSeek V4 Pro: 45.0 (#51), o3-pro: 72.2 (#1)
| Benchmark | DeepSeek V4 Pro | o3-pro |
|---|---|---|
| Fiction.LiveBench | — | 97.2% |
| CL-bench Life | 13.5% | — |
| LMArena Longer Query | 1458 | — |
Writing & Preference DeepSeek V4 Pro leads
DeepSeek V4 Pro: 65.5 (#46), o3-pro: 57.1 (#133)
| Benchmark | DeepSeek V4 Pro | o3-pro |
|---|---|---|
| LMArena Text | 1451 | — |
| LMArena Creative Writing | 1446 | — |
| Short-Story Creative Writing | — | 84.4% |
| EQ-Bench Creative Writing | 1553 | — |
| EQ-Bench 4 | 1166 | — |
| LMArena Multi-Turn | 1467 | — |
Frequently asked questions
Is DeepSeek V4 Pro better than o3-pro?
DeepSeek V4 Pro is the stronger model overall, scoring 54.3 to 42.9 on the Noometry Index.
Which is cheaper, DeepSeek V4 Pro or o3-pro?
DeepSeek V4 Pro is cheaper. It lists at $0.66 per million input tokens and $1.98 per million output tokens; o3-pro lists at $20 and $80.
Is DeepSeek V4 Pro or o3-pro better for coding?
o3-pro scores higher on coding benchmarks: 55.5 versus 52.4 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 o3-pro share?
8 benchmarks have published results for both models. DeepSeek V4 Pro has 48 scored results on Noometry and o3-pro has 12.