Model comparison
GPT-5 Mini vs Grok 4.7
Grok 4.7 is the stronger model overall, scoring 46.6 to 41.7 on the Noometry Index. GPT-5 Mini costs 4.4× less per token, which makes it the better buy when Grok 4.7's lead doesn't matter for your workload.
Last verified . 16 shared benchmarks.
Summary
- They share 16 benchmarks with published results for both. GPT-5 Mini scores higher in 4 categories and Grok 4.7 in 5 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Grok 4.7 leads 40.1 to 24.1.
- The biggest single-benchmark swing is ProofBench: 9% for GPT-5 Mini and 34% for Grok 4.7.
- GPT-5 Mini is cheaper at $0.25 / $2 per million input/output tokens, against $2 / $6 for Grok 4.7.
- Grok 4.7 accepts more context: 500K tokens versus 400K.
Side by side
| GPT-5 Mini | Grok 4.7 | |
|---|---|---|
| Provider | OpenAI | xAI |
| Noometry Index | 41.7 | 46.6 |
| Released | 2025-08-07 | 2026-09-21 |
| Weights | Proprietary | Proprietary |
| Context window | 400K | 500K |
| Max output | 128K | 500K |
| Input $ / M tokens | $0.25 | $2 |
| Output $ / M tokens | $2 | $6 |
| Results tracked | 60 | 21 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Grok 4.7 leads
GPT-5 Mini: 40.0 (#134), Grok 4.7: 55.0 (#23)
| Benchmark | GPT-5 Mini | Grok 4.7 |
|---|---|---|
| SciCode | 39.2% | 57.4% |
| LMArena Coding | 1406 | 1423 |
| SWE-bench Verified | 64.7% | — |
| SWE-bench Verified (bash only) | 59.8% | — |
| CursorBench | — | 46.3% |
| LMArena WebDev | — | 1636 |
| SWE-bench Multilingual | 39.7% | — |
| FrontierSWE | — | 29.5% |
| WeirdML | 52.7% | — |
| ALE-Bench | 799.77 | — |
| AlgoTune | 1.38 | — |
Agentic & Tool Use Not comparable
GPT-5 Mini: 31.0 (#64), Grok 4.7: —
| Benchmark | GPT-5 Mini | Grok 4.7 |
|---|---|---|
| Terminal-Bench | 34.8% | — |
| Berkeley Function Calling Leaderboard | 55.5% | — |
| Vending-Bench 2 | -31.18 | — |
Reasoning Grok 4.7 leads
GPT-5 Mini: 24.1 (#153), Grok 4.7: 40.1 (#49)
| Benchmark | GPT-5 Mini | Grok 4.7 |
|---|---|---|
| CritPt | 0% | 17.7% |
| LMArena Hard Prompts | 1380 | 1409 |
| ARC-AGI-2 | 4.4% | — |
| Kagi LLM Benchmark | 70.3% | — |
| NYT Connections (extended) | — | 76.8% |
| ARC-AGI-1 | 54.3% | — |
| Chess Puzzles | 30% | — |
| EnigmaEval | 8.2% | — |
| Mystery Game Puzzles | 10% | — |
| DTBench | 80.5% | — |
| LMCA | 34.2% | — |
| Epoch Capabilities Index | 145.51 | — |
| ForecastBench | 61 | — |
Math GPT-5 Mini leads
GPT-5 Mini: 46.3 (#59), Grok 4.7: 41.1 (#73)
| Benchmark | GPT-5 Mini | Grok 4.7 |
|---|---|---|
| ProofBench | 9% | 34% |
| LMArena Math | 1377 | 1393 |
| FrontierMath (Tiers 1-3) | 46.7% | — |
| FrontierMath Tier 4 | 12.2% | — |
| OTIS Mock AIME 2024-2025 | 86.7% | — |
| Omni-MATH | 72.2% | — |
| MATH Level 5 | 97.8% | — |
| FrontierMath (Feb 2025 set) | 27.2% | — |
| FrontierMath Tier 4 (v1) | 6.3% | — |
Knowledge GPT-5 Mini leads
GPT-5 Mini: 45.5 (#80), Grok 4.7: 38.7 (#130)
| Benchmark | GPT-5 Mini | Grok 4.7 |
|---|---|---|
| LMArena Expert | 1377 | 1413 |
| GPQA Diamond | 75% | — |
| Humanity's Last Exam | 19.4% | — |
| SimpleQA Verified | 21.6% | — |
| MMLU-Pro | 83.5% | — |
| Confabulations | 13.3% | — |
| Vectara Hallucination Rate | 12.9% | — |
| GPQA (HELM) | 75.6% | — |
Multimodal GPT-5 Mini leads
GPT-5 Mini: 35.0 (#80), Grok 4.7: 28.8 (#107)
| Benchmark | GPT-5 Mini | Grok 4.7 |
|---|---|---|
| LMArena Vision | 1202 | — |
| VPCT | 40.2% | — |
| Furniture Assembly | — | 20.8% |
Multilingual Grok 4.7 leads
GPT-5 Mini: 49.1 (#131), Grok 4.7: 51.2 (#108)
| Benchmark | GPT-5 Mini | Grok 4.7 |
|---|---|---|
| LMArena Non-English | 1363 | 1392 |
| LMArena Chinese | 1387 | 1476 |
| LMArena Russian | 1362 | 1396 |
| LMArena French | 1385 | — |
| LMArena German | 1366 | — |
| LMArena Japanese | 1342 | — |
| LMArena Korean | 1306 | — |
| LMArena Spanish | 1355 | — |
Instruction Following GPT-5 Mini leads
GPT-5 Mini: 76.4 (#43), Grok 4.7: 74.3 (#98)
| Benchmark | GPT-5 Mini | Grok 4.7 |
|---|---|---|
| LMArena Instruction Following | 1357 | 1406 |
| IFEval | 92.7% | — |
Long Context Grok 4.7 leads
GPT-5 Mini: 41.9 (#127), Grok 4.7: 43.2 (#97)
| Benchmark | GPT-5 Mini | Grok 4.7 |
|---|---|---|
| LMArena Longer Query | 1355 | 1416 |
| Fiction.LiveBench | 69.4% | — |
Writing & Preference Grok 4.7 leads
GPT-5 Mini: 55.4 (#143), Grok 4.7: 70.1 (#23)
| Benchmark | GPT-5 Mini | Grok 4.7 |
|---|---|---|
| LMArena Text | 1373 | 1397 |
| LMArena Creative Writing | 1325 | 1391 |
| EQ-Bench Creative Writing | 1313 | 2007 |
| LMArena Multi-Turn | 1363 | 1389 |
| Short-Story Creative Writing | 83.1% | — |
| WildBench | 85.5% | — |
Frequently asked questions
Is GPT-5 Mini better than Grok 4.7?
Grok 4.7 is the stronger model overall, scoring 46.6 to 41.7 on the Noometry Index. GPT-5 Mini costs 4.4× less per token, which makes it the better buy when Grok 4.7's lead doesn't matter for your workload.
Which is cheaper, GPT-5 Mini or Grok 4.7?
GPT-5 Mini is cheaper. It lists at $0.25 per million input tokens and $2 per million output tokens; Grok 4.7 lists at $2 and $6.
Is GPT-5 Mini or Grok 4.7 better for coding?
Grok 4.7 scores higher on coding benchmarks: 55.0 versus 40.0 in the Noometry coding category.
Which has the bigger context window?
Grok 4.7 does, with 500K tokens against 400K.
How many benchmarks do GPT-5 Mini and Grok 4.7 share?
16 benchmarks have published results for both models. GPT-5 Mini has 60 scored results on Noometry and Grok 4.7 has 21.