Model comparison
DeepSeek V4 Pro vs Grok 4.5
DeepSeek V4 Pro and Grok 4.5 score almost the same on the Noometry Index (54.3 vs 55.0), so choose on price, context window or the category you care about most.
Last verified . 41 shared benchmarks.
DeepSeekSummary
- They share 41 benchmarks with published results for both. DeepSeek V4 Pro scores higher in 5 categories and Grok 4.5 in 4 categories; 3 gaps are clear of the uncertainty.
- The widest gap is in agentic & tool use, where Grok 4.5 leads 44.4 to 32.8.
- The biggest single-benchmark swing is Surface Evolver Bench: 40% for DeepSeek V4 Pro and 74.4% for Grok 4.5.
- DeepSeek V4 Pro is cheaper at $0.66 / $1.98 per million input/output tokens, against $2 / $6 for Grok 4.5.
- DeepSeek V4 Pro accepts more context: 1M tokens versus 500K.
- DeepSeek V4 Pro has downloadable open weights; the other is API-only.
Side by side
| DeepSeek V4 Pro | Grok 4.5 | |
|---|---|---|
| Provider | DeepSeek | xAI |
| Noometry Index | 54.3 | 55.0 |
| Released | 2026-04-24 | 2026-07-08 |
| Weights | Open | Proprietary |
| Context window | 1M | 500K |
| Max output | 393K | 500K |
| Input $ / M tokens | $0.66 | $2 |
| Output $ / M tokens | $1.98 | $6 |
| Results tracked | 48 | 52 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
DeepSeek V4 Pro: 52.4 (#34), Grok 4.5: 52.2 (#35)
| Benchmark | DeepSeek V4 Pro | Grok 4.5 |
|---|---|---|
| FrontierCode | 28.6% | 42.4% |
| LMArena WebDev | 1582 | 1553 |
| SciCode | 51% | 54.1% |
| WeirdML | 66.2% | 46.4% |
| LMArena Coding | 1470 | 1474 |
| ALE-Bench | 1,403 | 1,309 |
| SWE-bench Verified | 77.6% | — |
| DeepSWE | — | 53.8% |
Agentic & Tool Use Grok 4.5 leads
DeepSeek V4 Pro: 32.8 (#58), Grok 4.5: 44.4 (#17)
| Benchmark | DeepSeek V4 Pro | Grok 4.5 |
|---|---|---|
| APEX-Agents | 47.3% | 56.2% |
| Vending-Bench 2 | 3,285 | 3,887 |
| τ²-bench Banking | — | 47.9% |
| PostTrainBench | — | 23.4% |
| GBAEval | — | 65.4% |
| GDP.pdf | — | 14% |
| LMArena Search | — | 1213 |
Reasoning Too close to call
DeepSeek V4 Pro: 56.5 (#24), Grok 4.5: 56.1 (#25)
| Benchmark | DeepSeek V4 Pro | Grok 4.5 |
|---|---|---|
| ARC-AGI-2 | 61.3% | 52.6% |
| Kagi LLM Benchmark | 53.5% | 83.5% |
| NYT Connections (extended) | 91.3% | 79.9% |
| ARC-AGI-1 | 90.5% | 87.2% |
| CritPt | 18% | 15.4% |
| Chess Puzzles | 47% | 36% |
| LMArena Hard Prompts | 1461 | 1462 |
| DTBench | 93.9% | 96.5% |
| LMCA | 45.5% | 45.2% |
| Surface Evolver Bench | 40% | 74.4% |
| Epoch Capabilities Index | 155.31 | 153.92 |
| SimpleBench | — | 70% |
| Mystery Game Puzzles | 43% | — |
| ForecastBench | 56.1 | — |
Math DeepSeek V4 Pro leads
DeepSeek V4 Pro: 64.8 (#30), Grok 4.5: 60.9 (#35)
| Benchmark | DeepSeek V4 Pro | Grok 4.5 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 64.6% | 57.2% |
| FrontierMath Tier 4 | 26.8% | 24.4% |
| OTIS Mock AIME 2024-2025 | 98.6% | 97.8% |
| ProofBench | 50% | 31% |
| LMArena Math | 1455 | 1459 |
| MathArena Final-Answer Competitions | 76.6% | — |
Knowledge Grok 4.5 leads
DeepSeek V4 Pro: 59.5 (#31), Grok 4.5: 62.3 (#24)
| Benchmark | DeepSeek V4 Pro | Grok 4.5 |
|---|---|---|
| GPQA Diamond | 91.7% | 93.4% |
| SimpleQA Verified | 52.9% | 48.3% |
| LMArena Expert | 1464 | 1466 |
| Vectara Hallucination Rate | 8.6% | — |
Multimodal Not comparable
DeepSeek V4 Pro: —, Grok 4.5: 37.6 (#72)
| Benchmark | DeepSeek V4 Pro | Grok 4.5 |
|---|---|---|
| LMArena Vision | — | 1288 |
| Blueprint-Bench 2 | — | 27.3% |
| Furniture Assembly | — | 22.5% |
| LMArena Document | — | 1452 |
Multilingual Too close to call
DeepSeek V4 Pro: 54.4 (#45), Grok 4.5: 54.4 (#42)
| Benchmark | DeepSeek V4 Pro | Grok 4.5 |
|---|---|---|
| LMArena Non-English | 1439 | 1440 |
| LMArena Chinese | 1486 | 1496 |
| LMArena French | 1472 | 1456 |
| LMArena German | 1458 | 1446 |
| LMArena Japanese | 1445 | 1428 |
| LMArena Korean | 1447 | 1404 |
| LMArena Russian | 1453 | 1448 |
| LMArena Spanish | 1458 | 1450 |
Instruction Following Too close to call
DeepSeek V4 Pro: 76.1 (#47), Grok 4.5: 76.0 (#48)
| Benchmark | DeepSeek V4 Pro | Grok 4.5 |
|---|---|---|
| LMArena Instruction Following | 1448 | 1446 |
Long Context Too close to call
DeepSeek V4 Pro: 45.0 (#51), Grok 4.5: 44.8 (#56)
| Benchmark | DeepSeek V4 Pro | Grok 4.5 |
|---|---|---|
| LMArena Longer Query | 1458 | 1463 |
| CL-bench Life | 13.5% | — |
Writing & Preference Too close to call
DeepSeek V4 Pro: 65.5 (#46), Grok 4.5: 65.8 (#42)
| Benchmark | DeepSeek V4 Pro | Grok 4.5 |
|---|---|---|
| LMArena Text | 1451 | 1448 |
| LMArena Creative Writing | 1446 | 1442 |
| EQ-Bench Creative Writing | 1553 | 1579 |
| LMArena Multi-Turn | 1467 | 1456 |
| EQ-Bench 4 | 1166 | — |
Frequently asked questions
Is DeepSeek V4 Pro better than Grok 4.5?
DeepSeek V4 Pro and Grok 4.5 score almost the same on the Noometry Index (54.3 vs 55.0), so choose on price, context window or the category you care about most.
Which is cheaper, DeepSeek V4 Pro or Grok 4.5?
DeepSeek V4 Pro is cheaper. It lists at $0.66 per million input tokens and $1.98 per million output tokens; Grok 4.5 lists at $2 and $6.
Is DeepSeek V4 Pro or Grok 4.5 better for coding?
They score almost the same on coding (52.4 vs 52.2); test both on your own repository before choosing.
Which has the bigger context window?
DeepSeek V4 Pro does, with 1M tokens against 500K.
How many benchmarks do DeepSeek V4 Pro and Grok 4.5 share?
41 benchmarks have published results for both models. DeepSeek V4 Pro has 48 scored results on Noometry and Grok 4.5 has 52.