Model comparison
DeepSeek-V3 vs Grok 4.7
Grok 4.7 is the stronger model overall, scoring 53.1 to 39.5 on the Noometry Index. DeepSeek-V3 costs 7.4× less per token, which makes it the better buy when Grok 4.7's lead doesn't matter for your workload.
Last verified . 22 shared benchmarks.
DeepSeekSummary
- They share 22 benchmarks with published results for both. DeepSeek-V3 scores higher in 0 categories and Grok 4.7 in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Grok 4.7 leads 49.1 to 20.5.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 37.8% for DeepSeek-V3 and 98.1% for Grok 4.7.
- DeepSeek-V3 is cheaper at $0.24 / $0.90 per million input/output tokens, against $2 / $6 for Grok 4.7.
- Grok 4.7 accepts more context: 500K tokens versus 164K.
- DeepSeek-V3 has downloadable open weights; the other is API-only.
Side by side
| DeepSeek-V3 | Grok 4.7 | |
|---|---|---|
| Provider | DeepSeek | xAI |
| Noometry Index | 39.5 | 53.1 |
| Released | 2024-12-26 | 2026-09-21 |
| Weights | Open | Proprietary |
| Context window | 164K | 500K |
| Max output | 164K | 500K |
| Input $ / M tokens | $0.24 | $2 |
| Output $ / M tokens | $0.90 | $6 |
| Results tracked | 60 | 39 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Grok 4.7 leads
DeepSeek-V3: 42.3 (#106), Grok 4.7: 58.0 (#18)
| Benchmark | DeepSeek-V3 | Grok 4.7 |
|---|---|---|
| SciCode | 35.8% | 57.8% |
| LMArena Coding | 1368 | 1427 |
| FrontierCode | — | 47.6% |
| Aider Polyglot | 55.1% | — |
| CursorBench | — | 46.3% |
| LMArena WebDev | — | 1639 |
| FrontierSWE | — | 29.5% |
| WeirdML | 36.1% | — |
| BigCodeBench Instruct | 50% | — |
| LiveBench Coding | 70.9% | — |
| BigCodeBench Complete | 62.2% | — |
| HumanEval+ | 86.6% | — |
| MBPP+ | 73% | — |
Agentic & Tool Use Not comparable
DeepSeek-V3: —, Grok 4.7: 36.7 (#37)
| Benchmark | DeepSeek-V3 | Grok 4.7 |
|---|---|---|
| APEX-Agents | — | 54.6% |
| GDP.pdf | — | 22.8% |
| METR Time Horizons | 49.6% | — |
| Vending-Bench 2 | — | 10,537 |
Reasoning Grok 4.7 leads
DeepSeek-V3: 20.5 (#236), Grok 4.7: 49.1 (#40)
| Benchmark | DeepSeek-V3 | Grok 4.7 |
|---|---|---|
| CritPt | 0% | 18% |
| LMArena Hard Prompts | 1365 | 1413 |
| DTBench | 64.8% | 96% |
| LMCA | 15.5% | 49.4% |
| Epoch Capabilities Index | 135.94 | 153.53 |
| SimpleBench | 27.2% | — |
| Kagi LLM Benchmark | 52.3% | — |
| NYT Connections (extended) | — | 76.8% |
| Chess Puzzles | — | 38% |
| LiveBench Reasoning | 65.8% | — |
| Mystery Game Puzzles | — | 29% |
| LiveBench Data Analysis | 60.9% | — |
| BIG-Bench Hard | 87.5% | — |
| ForecastBench | 59.1 | — |
| HellaSwag | 88.9% | — |
| LiveBench | 66.9% | — |
| PIQA | 84.7% | — |
| WinoGrande | 85.2% | — |
Math Grok 4.7 leads
DeepSeek-V3: 32.1 (#219), Grok 4.7: 57.8 (#39)
| Benchmark | DeepSeek-V3 | Grok 4.7 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 37.8% | 98.1% |
| LMArena Math | 1373 | 1407 |
| FrontierMath (Tiers 1-3) | — | 53% |
| FrontierMath Tier 4 | — | 17.1% |
| ProofBench | — | 34% |
| Omni-MATH | 40.3% | — |
| LiveBench Math | 73.5% | — |
| MATH Level 5 | 75.5% | — |
| FrontierMath (Feb 2025 set) | 1.7% | — |
Knowledge Grok 4.7 leads
DeepSeek-V3: 37.5 (#155), Grok 4.7: 62.8 (#22)
| Benchmark | DeepSeek-V3 | Grok 4.7 |
|---|---|---|
| GPQA Diamond | 67.6% | 92.7% |
| LMArena Expert | 1351 | 1422 |
| SimpleQA Verified | — | 56% |
| MMLU-Pro | 72.3% | — |
| Confabulations | 26.1% | — |
| Vectara Hallucination Rate | 6.1% | — |
| GPQA (HELM) | 53.8% | — |
| ARC (AI2) Challenge | 95.3% | — |
| MMLU | 87.2% | — |
| TriviaQA | 82.9% | — |
Multimodal Not comparable
DeepSeek-V3: —, Grok 4.7: 35.5 (#87)
| Benchmark | DeepSeek-V3 | Grok 4.7 |
|---|---|---|
| LMArena Vision | — | 1228 |
| Blueprint-Bench 2 | — | 32.5% |
| Furniture Assembly | — | 20.8% |
Multilingual Grok 4.7 leads
DeepSeek-V3: 48.5 (#143), Grok 4.7: 50.8 (#116)
| Benchmark | DeepSeek-V3 | Grok 4.7 |
|---|---|---|
| LMArena Non-English | 1358 | 1389 |
| LMArena Chinese | 1391 | 1455 |
| LMArena French | 1385 | 1455 |
| LMArena Russian | 1373 | 1397 |
| LMArena Spanish | 1358 | 1400 |
| LMArena German | 1374 | — |
| LMArena Japanese | 1333 | — |
| LMArena Korean | 1319 | — |
Instruction Following Grok 4.7 leads
DeepSeek-V3: 72.8 (#130), Grok 4.7: 74.1 (#105)
| Benchmark | DeepSeek-V3 | Grok 4.7 |
|---|---|---|
| LMArena Instruction Following | 1345 | 1404 |
| LiveBench Instruction Following | 81.5% | — |
| IFEval | 83.2% | — |
Long Context Grok 4.7 leads
DeepSeek-V3: 34.0 (#253), Grok 4.7: 43.1 (#104)
| Benchmark | DeepSeek-V3 | Grok 4.7 |
|---|---|---|
| LMArena Longer Query | 1352 | 1413 |
| Fiction.LiveBench | 50% | — |
Writing & Preference Grok 4.7 leads
DeepSeek-V3: 57.4 (#130), Grok 4.7: 70.0 (#24)
| Benchmark | DeepSeek-V3 | Grok 4.7 |
|---|---|---|
| LMArena Text | 1375 | 1399 |
| LMArena Creative Writing | 1364 | 1391 |
| EQ-Bench Creative Writing | 1472 | 2007 |
| LMArena Multi-Turn | 1389 | 1393 |
| Short-Story Creative Writing | 77% | — |
| WildBench | 83% | — |
| LiveBench Language | 49.1% | — |
Frequently asked questions
Is DeepSeek-V3 better than Grok 4.7?
Grok 4.7 is the stronger model overall, scoring 53.1 to 39.5 on the Noometry Index. DeepSeek-V3 costs 7.4× less per token, which makes it the better buy when Grok 4.7's lead doesn't matter for your workload.
Which is cheaper, DeepSeek-V3 or Grok 4.7?
DeepSeek-V3 is cheaper. It lists at $0.24 per million input tokens and $0.90 per million output tokens; Grok 4.7 lists at $2 and $6.
Is DeepSeek-V3 or Grok 4.7 better for coding?
Grok 4.7 scores higher on coding benchmarks: 58.0 versus 42.3 in the Noometry coding category.
Which has the bigger context window?
Grok 4.7 does, with 500K tokens against 164K.
How many benchmarks do DeepSeek-V3 and Grok 4.7 share?
22 benchmarks have published results for both models. DeepSeek-V3 has 60 scored results on Noometry and Grok 4.7 has 39.