Model comparison
GPT-4.1 vs Grok 4.5
Grok 4.5 is the stronger model overall, scoring 55.0 to 35.9 on the Noometry Index.
Last verified . 33 shared benchmarks.
Summary
- They share 33 benchmarks with published results for both. GPT-4.1 scores higher in 1 category and Grok 4.5 in 9 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Grok 4.5 leads 56.1 to 11.7.
- The biggest single-benchmark swing is ARC-AGI-1: 5.5% for GPT-4.1 and 87.2% for Grok 4.5.
- Grok 4.5 is cheaper at $2 / $6 per million input/output tokens, against $2 / $8 for GPT-4.1.
- GPT-4.1 accepts more context: 1.05M tokens versus 500K.
Side by side
| GPT-4.1 | Grok 4.5 | |
|---|---|---|
| Provider | OpenAI | xAI |
| Noometry Index | 35.9 | 55.0 |
| Released | 2025-04-14 | 2026-07-08 |
| Weights | Proprietary | Proprietary |
| Context window | 1.05M | 500K |
| Max output | 33K | 500K |
| Input $ / M tokens | $2 | $2 |
| Output $ / M tokens | $8 | $6 |
| Results tracked | 52 | 52 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Grok 4.5 leads
GPT-4.1: 34.4 (#238), Grok 4.5: 52.2 (#35)
| Benchmark | GPT-4.1 | Grok 4.5 |
|---|---|---|
| WeirdML | 39% | 46.4% |
| LMArena Coding | 1391 | 1474 |
| ALE-Bench | 558.1 | 1,309 |
| SWE-bench Verified | 48.5% | — |
| DeepSWE | — | 53.8% |
| FrontierCode | — | 42.4% |
| SWE-bench Verified (bash only) | 39.6% | — |
| Aider Polyglot | 52.4% | — |
| LMArena WebDev | — | 1553 |
| SciCode | — | 54.1% |
| CadEval | 42% | — |
Agentic & Tool Use Grok 4.5 leads
GPT-4.1: 34.7 (#43), Grok 4.5: 44.4 (#17)
| Benchmark | GPT-4.1 | Grok 4.5 |
|---|---|---|
| APEX-Agents | — | 56.2% |
| Berkeley Function Calling Leaderboard | 54% | — |
| τ²-bench Banking | — | 47.9% |
| PostTrainBench | — | 23.4% |
| GBAEval | — | 65.4% |
| GDP.pdf | — | 14% |
| LMArena Search | — | 1213 |
| Vending-Bench 2 | — | 3,887 |
Reasoning Grok 4.5 leads
GPT-4.1: 11.7 (#339), Grok 4.5: 56.1 (#25)
| Benchmark | GPT-4.1 | Grok 4.5 |
|---|---|---|
| ARC-AGI-2 | 0.4% | 52.6% |
| SimpleBench | 27% | 70% |
| Kagi LLM Benchmark | 52.3% | 83.5% |
| ARC-AGI-1 | 5.5% | 87.2% |
| Chess Puzzles | 6% | 36% |
| LMArena Hard Prompts | 1384 | 1462 |
| DTBench | 68.3% | 96.5% |
| LMCA | 25.6% | 45.2% |
| Epoch Capabilities Index | 136.78 | 153.92 |
| NYT Connections (extended) | — | 79.9% |
| CritPt | — | 15.4% |
| EnigmaEval | 2.2% | — |
| Surface Evolver Bench | — | 74.4% |
| ForecastBench | 61.5 | — |
Math Grok 4.5 leads
GPT-4.1: 22.3 (#280), Grok 4.5: 60.9 (#35)
| Benchmark | GPT-4.1 | Grok 4.5 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 6% | 57.2% |
| OTIS Mock AIME 2024-2025 | 38.3% | 97.8% |
| LMArena Math | 1370 | 1459 |
| FrontierMath Tier 4 | — | 24.4% |
| ProofBench | — | 31% |
| Omni-MATH | 47.1% | — |
| MATH Level 5 | 83% | — |
| FrontierMath (Feb 2025 set) | 5.5% | — |
| FrontierMath Tier 4 (v1) | 0% | — |
Knowledge Grok 4.5 leads
GPT-4.1: 37.1 (#160), Grok 4.5: 62.3 (#24)
| Benchmark | GPT-4.1 | Grok 4.5 |
|---|---|---|
| GPQA Diamond | 66.9% | 93.4% |
| SimpleQA Verified | 31.1% | 48.3% |
| LMArena Expert | 1364 | 1466 |
| Humanity's Last Exam | 5.4% | — |
| MMLU-Pro | 81.1% | — |
| Vectara Hallucination Rate | 5.6% | — |
| GPQA (HELM) | 65.9% | — |
Multimodal Too close to call
GPT-4.1: 38.2 (#67), Grok 4.5: 37.6 (#72)
| Benchmark | GPT-4.1 | Grok 4.5 |
|---|---|---|
| LMArena Vision | 1211 | 1288 |
| GeoBench | 72% | — |
| Blueprint-Bench 2 | — | 27.3% |
| Furniture Assembly | — | 22.5% |
| LMArena Document | — | 1452 |
Multilingual Grok 4.5 leads
GPT-4.1: 49.4 (#133), Grok 4.5: 54.4 (#42)
| Benchmark | GPT-4.1 | Grok 4.5 |
|---|---|---|
| LMArena Non-English | 1370 | 1440 |
| LMArena Chinese | 1382 | 1496 |
| LMArena French | 1382 | 1456 |
| LMArena German | 1381 | 1446 |
| LMArena Japanese | 1319 | 1428 |
| LMArena Korean | 1339 | 1404 |
| LMArena Russian | 1377 | 1448 |
| LMArena Spanish | 1376 | 1450 |
Instruction Following Grok 4.5 leads
GPT-4.1: 71.3 (#153), Grok 4.5: 76.0 (#48)
| Benchmark | GPT-4.1 | Grok 4.5 |
|---|---|---|
| LMArena Instruction Following | 1367 | 1446 |
| IFEval | 83.8% | — |
Long Context Grok 4.5 leads
GPT-4.1: 40.0 (#163), Grok 4.5: 44.8 (#56)
| Benchmark | GPT-4.1 | Grok 4.5 |
|---|---|---|
| LMArena Longer Query | 1385 | 1463 |
| Fiction.LiveBench | 63.9% | — |
Writing & Preference Grok 4.5 leads
GPT-4.1: 57.6 (#125), Grok 4.5: 65.8 (#42)
| Benchmark | GPT-4.1 | Grok 4.5 |
|---|---|---|
| LMArena Text | 1383 | 1448 |
| LMArena Creative Writing | 1363 | 1442 |
| EQ-Bench Creative Writing | 1420 | 1579 |
| LMArena Multi-Turn | 1398 | 1456 |
| WildBench | 85.4% | — |
Frequently asked questions
Is GPT-4.1 better than Grok 4.5?
Grok 4.5 is the stronger model overall, scoring 55.0 to 35.9 on the Noometry Index.
Which is cheaper, GPT-4.1 or Grok 4.5?
Grok 4.5 is cheaper. It lists at $2 per million input tokens and $6 per million output tokens; GPT-4.1 lists at $2 and $8.
Is GPT-4.1 or Grok 4.5 better for coding?
Grok 4.5 scores higher on coding benchmarks: 52.2 versus 34.4 in the Noometry coding category.
Which has the bigger context window?
GPT-4.1 does, with 1.05M tokens against 500K.
How many benchmarks do GPT-4.1 and Grok 4.5 share?
33 benchmarks have published results for both models. GPT-4.1 has 52 scored results on Noometry and Grok 4.5 has 52.