Grok 4 vs Grok 4.6
Grok 4 Verdict
- Grok 4.6 leads on quality: 62.7 vs 59.8.
- Grok 4 is stronger in human preference, instruction following, long context, multimodal.
- Grok 4.6 is stronger in agents & tools, coding, knowledge, maths, reasoning, composite.
- Grok 4 is 1.9× cheaper ($1.56 vs $3.00 per 1M blended).
| Metric | Grok 4 | Grok 4.6 |
|---|---|---|
| BenchLeader Index | 59.8 | 62.7 |
| Agents & tools score | 53.4 | 68.6 |
| Coding score | 61.2 | 64.5 |
| Human preference score | 59.8 | – |
| Instruction following score | 67.7 | – |
| Knowledge score | 56.0 | 75.3 |
| Long context score | 77.7 | 64.3 |
| Maths score | 58.0 | 58.6 |
| Multimodal score | 54.5 | – |
| Reasoning score | 58.3 | 71.8 |
| Composite score | – | 69.3 |
| Blended price $/M | $1.56 | $3.00 |
| Output speed | – | 58 tok/s |
| Time to first answer | – | 31.3 s |
| Context window | 256k | 500k |
| GPQA Diamond | 87.0% | – |
| OTIS Mock AIME | 84.0% | – |
| Terminal-Bench | 27.2% | – |
| SimpleBench | 60.5% | 75.9% |
| Fiction.LiveBench 120k | 96.9% | – |
| Cybench | 43.0% | – |
| WeirdML | 45.7% | – |
| APEX-Agents | 15.2% | 41.2% |
| FrontierCode | – | 48.0% |
| ProofBench | – | 51.0% |
| Epoch Capabilities Index | 146.4 | 156.3 |
| LMArena Text | 1411 | – |
| LMArena Hard Prompts | 1420 | – |
| LMArena Coding | 1435 | – |
| LMArena Vision | 1210 | – |
| LiveBench | – | 78.0% |
| LiveBench Reasoning | – | 90.5% |
| LiveBench Coding | – | 76.8% |
| LiveBench Agentic Coding | – | 57.0% |
| LiveBench Mathematics | – | 92.6% |
| LiveBench Data Analysis | – | 73.9% |
| LiveBench Language | – | 83.7% |
| AA Intelligence Index | – | 44.4 |
| AA-LCR | – | 80.3% |
| AA-Omniscience | – | 30.5 |
| GPQA Diamond (AA) | – | 95.0% |
| Humanity's Last Exam (AA) | – | 42.9% |
| SciCode (AA) | – | 56.5% |
| AIME (Vals) | 90.6% | – |
| LiveCodeBench | 83.3% | – |
| MMLU-Pro | 85.3% | – |
| LegalBench | 83.2% | – |
| CorpFin | 66.0% | – |
| TaxEval | 65.1% | – |
| MedQA | 92.5% | – |
| MGSM | 90.9% | – |
| SWE-bench (Vals) | 57.8% | – |
| GPQA Diamond (Vals) | 88.1% | – |
| IMO 2025 | 21.4% | – |
| MathArena Apex | 2.1% | – |
| Kagi LLM Benchmark | 73.6% | – |
| IFEval (HELM) | 94.9% | – |
| Omni-MATH (HELM) | 60.3% | – |
| WildBench (HELM) | 79.7% | – |
| MMLU-Pro (HELM) | 85.1% | – |
| GPQA Diamond (HELM) | 72.6% | – |
| HELM Capabilities mean | 78.5% | – |
| Aider Polyglot | 79.6% | – |
| ARC-AGI-1 | 79.6% | – |
| ARC-AGI-2 | 29.4% | – |
| BFCL Overall | 63.0% | – |
Data as of 2026-09-13. Best configuration of each model; every score links to its source on the model pages.
Grok 4 vs Grok 4.6: questions
- Is Grok 4 better than Grok 4.6?
- Grok 4.6 leads on quality: 62.7 vs 59.8. The BenchLeader Index combines every independent quality benchmark; Grok 4.6 is ahead overall as of 2026-09-13, but check the category scores for your use.
- Is Grok 4 better than Grok 4.6 for coding?
- Grok 4.6 scores higher in coding (65 vs 61 on the category index, where 50 is average).
- Is Grok 4 better than Grok 4.6 for agentic tasks?
- Grok 4.6 scores higher in agentic tasks (69 vs 53 on the category index, where 50 is average).
- Which is cheaper, Grok 4 or Grok 4.6?
- Grok 4 is cheaper: $1.56 against $3.00 per million tokens, blended at three input tokens per output token.
- Which has the larger context window?
- Grok 4.6 accepts more context: 500k against 256k tokens.