Model comparison
DeepSeek-R1 vs Grok 4.3
Grok 4.3 is the stronger model overall, scoring 43.8 to 42.3 on the Noometry Index. DeepSeek-R1 costs 1.7× less per token, which makes it the better buy when Grok 4.3's lead doesn't matter for your workload.
Last verified . 25 shared benchmarks.
DeepSeekSummary
- They share 25 benchmarks with published results for both. DeepSeek-R1 scores higher in 5 categories and Grok 4.3 in 4 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Grok 4.3 leads 35.9 to 18.6.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 66.4% for DeepSeek-R1 and 93.3% for Grok 4.3.
- DeepSeek-R1 is cheaper at $0.50 / $2.15 per million input/output tokens, against $1.25 / $2.50 for Grok 4.3.
- Grok 4.3 accepts more context: 1M tokens versus 164K.
Side by side
| DeepSeek-R1 | Grok 4.3 | |
|---|---|---|
| Provider | DeepSeek | xAI |
| Noometry Index | 42.3 | 43.8 |
| Released | 2025-01-20 | 2026-04-17 |
| Weights | Proprietary | Proprietary |
| Context window | 164K | 1M |
| Max output | 64K | 30K |
| Input $ / M tokens | $0.50 | $1.25 |
| Output $ / M tokens | $2.15 | $2.50 |
| Results tracked | 52 | 40 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek-R1 leads
DeepSeek-R1: 46.3 (#68), Grok 4.3: 41.6 (#121)
| Benchmark | DeepSeek-R1 | Grok 4.3 |
|---|---|---|
| SciCode | 35.7% | 47.3% |
| WeirdML | 41.6% | 49.9% |
| LMArena Coding | 1427 | 1415 |
| ALE-Bench | 804.12 | 944.17 |
| Aider Polyglot | 71.4% | — |
| LMArena WebDev | — | 1357 |
| LiveBench Coding | 66.7% | — |
| AlgoTune | 1.7 | — |
Agentic & Tool Use DeepSeek-R1 leads
DeepSeek-R1: 30.7 (#75), Grok 4.3: 27.7 (#99)
| Benchmark | DeepSeek-R1 | Grok 4.3 |
|---|---|---|
| DeepResearch Bench | 35.1% | — |
| BALROG | 34.9% | — |
| GDP.pdf | — | 8% |
| LMArena Search | — | 1165 |
| METR Time Horizons | 53.8% | — |
| Vending-Bench 2 | — | 35.26 |
Reasoning Grok 4.3 leads
DeepSeek-R1: 18.6 (#278), Grok 4.3: 35.9 (#68)
| Benchmark | DeepSeek-R1 | Grok 4.3 |
|---|---|---|
| CritPt | 1.1% | 8% |
| LMArena Hard Prompts | 1416 | 1396 |
| Epoch Capabilities Index | 141.29 | 149.16 |
| ForecastBench | 60 | 60.3 |
| ARC-AGI-2 | 1.3% | — |
| SimpleBench | 40.8% | — |
| Kagi LLM Benchmark | 69.4% | — |
| NYT Connections (extended) | — | 55.2% |
| ARC-AGI-1 | 21.2% | — |
| Chess Puzzles | — | 25% |
| LiveBench Reasoning | 83.2% | — |
| DTBench | — | 90.7% |
| LiveBench Data Analysis | 69.8% | — |
| LMCA | — | 38.3% |
| LiveBench | 71.6% | — |
Math Grok 4.3 leads
DeepSeek-R1: 43.8 (#79), Grok 4.3: 46.0 (#74)
| Benchmark | DeepSeek-R1 | Grok 4.3 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 66.4% | 93.3% |
| LMArena Math | 1400 | 1388 |
| FrontierMath (Tiers 1-3) | — | 42.8% |
| FrontierMath Tier 4 | — | 14.6% |
| ProofBench | — | 11% |
| Omni-MATH | 42.4% | — |
| LiveBench Math | 80.7% | — |
| MATH Level 5 | 96.6% | — |
Knowledge Grok 4.3 leads
DeepSeek-R1: 44.5 (#87), Grok 4.3: 52.5 (#62)
| Benchmark | DeepSeek-R1 | Grok 4.3 |
|---|---|---|
| GPQA Diamond | 76.3% | 88.8% |
| LMArena Expert | 1394 | 1385 |
| SimpleQA Verified | — | 33.2% |
| MMLU-Pro | 79.3% | — |
| Confabulations | 12.7% | — |
| Vectara Hallucination Rate | 11.3% | — |
| GPQA (HELM) | 66.6% | — |
Multimodal Not comparable
DeepSeek-R1: —, Grok 4.3: 31.6 (#104)
| Benchmark | DeepSeek-R1 | Grok 4.3 |
|---|---|---|
| LMArena Vision | — | 1229 |
| Blueprint-Bench 2 | — | 0% |
Multilingual DeepSeek-R1 leads
DeepSeek-R1: 52.4 (#85), Grok 4.3: 50.5 (#120)
| Benchmark | DeepSeek-R1 | Grok 4.3 |
|---|---|---|
| LMArena Non-English | 1412 | 1385 |
| LMArena Chinese | 1442 | 1422 |
| LMArena French | 1417 | 1412 |
| LMArena German | 1404 | 1395 |
| LMArena Japanese | 1391 | 1379 |
| LMArena Korean | 1360 | 1356 |
| LMArena Russian | 1423 | 1399 |
| LMArena Spanish | 1411 | 1398 |
Instruction Following Too close to call
DeepSeek-R1: 72.0 (#143), Grok 4.3: 72.1 (#140)
| Benchmark | DeepSeek-R1 | Grok 4.3 |
|---|---|---|
| LMArena Instruction Following | 1382 | 1366 |
| LiveBench Instruction Following | 80.5% | — |
| IFEval | 78.4% | — |
Long Context DeepSeek-R1 leads
DeepSeek-R1: 45.4 (#36), Grok 4.3: 42.5 (#123)
| Benchmark | DeepSeek-R1 | Grok 4.3 |
|---|---|---|
| LMArena Longer Query | 1391 | 1393 |
| Fiction.LiveBench | 75% | — |
Writing & Preference DeepSeek-R1 leads
DeepSeek-R1: 61.4 (#88), Grok 4.3: 58.5 (#118)
| Benchmark | DeepSeek-R1 | Grok 4.3 |
|---|---|---|
| LMArena Text | 1428 | 1397 |
| LMArena Creative Writing | 1405 | 1380 |
| LMArena Multi-Turn | 1405 | 1406 |
| Short-Story Creative Writing | 83% | — |
| EQ-Bench Creative Writing | 1500 | — |
| WildBench | 82.8% | — |
| EQ-Bench 4 | — | 1075 |
| LiveBench Language | 48.5% | — |
Frequently asked questions
Is DeepSeek-R1 better than Grok 4.3?
Grok 4.3 is the stronger model overall, scoring 43.8 to 42.3 on the Noometry Index. DeepSeek-R1 costs 1.7× less per token, which makes it the better buy when Grok 4.3's lead doesn't matter for your workload.
Which is cheaper, DeepSeek-R1 or Grok 4.3?
DeepSeek-R1 is cheaper. It lists at $0.50 per million input tokens and $2.15 per million output tokens; Grok 4.3 lists at $1.25 and $2.50.
Is DeepSeek-R1 or Grok 4.3 better for coding?
DeepSeek-R1 scores higher on coding benchmarks: 46.3 versus 41.6 in the Noometry coding category.
Which has the bigger context window?
Grok 4.3 does, with 1M tokens against 164K.
How many benchmarks do DeepSeek-R1 and Grok 4.3 share?
25 benchmarks have published results for both models. DeepSeek-R1 has 52 scored results on Noometry and Grok 4.3 has 40.