Claude Opus 5.5 vs Grok 4.20
Claude Opus 5.5 
Grok 4.20 Verdict
- Claude Opus 5.5 (thinking) leads on quality: 70.7 vs 59.7.
- Claude Opus 5.5 (thinking) is stronger in composite, knowledge, long context, multimodal, reasoning.
- Grok 4.20 (thinking) is stronger in agents & tools, coding, human preference, instruction following, maths.
- Grok 4.20 (thinking) is 5.1× cheaper ($1.56 vs $8.00 per 1M blended).
- Grok 4.20 (thinking) streams 2.1× faster (103 vs 48 tokens per second).
| Metric | Claude Opus 5.5 (thinking) | Grok 4.20 (thinking) |
|---|---|---|
| BenchLeader Index | 70.7 | 59.7 |
| Composite score | 95.0 | 60.5 |
| Knowledge score | 85.0 | 56.7 |
| Long context score | 68.5 | 60.4 |
| Multimodal score | 71.9 | 59.7 |
| Reasoning score | 95.0 | 60.5 |
| Agents & tools score | – | 51.6 |
| Coding score | – | 54.7 |
| Human preference score | – | 67.2 |
| Instruction following score | – | 79.9 |
| Maths score | – | 55.5 |
| Blended price $/M | $8.00 | $1.56 |
| Output speed | 48 tok/s | 103 tok/s |
| Time to first answer | 4.2 s | 25.9 s |
| Context window | 1M | 2M |
| GPQA Diamond | – | 89.3% |
| FrontierMath Tiers 1–3 | – | 44.9% |
| FrontierMath Tier 4 | – | 17.1% |
| OTIS Mock AIME | – | 92.2% |
| SimpleQA Verified | – | 30.2% |
| ProofBench | – | 14.0% |
| LMArena Text | – | 1472 |
| LMArena Hard Prompts | – | 1488 |
| LMArena Coding | – | 1510 |
| LMArena WebDev | – | 1374 |
| LMArena Vision | – | 1263 |
| AA Intelligence Index | 57.6 | 25.7 |
| IFBench | – | 82.9% |
| AA-LCR | 84.7% | 69.0% |
| MMMU-Pro | 87.7% | 74.6% |
| AA-Omniscience | 46.4 | 14.8 |
| Terminal-Bench Hard | – | 40.9% |
| GPQA Diamond (AA) | – | 91.1% |
| Humanity's Last Exam (AA) | 61.4% | 34.5% |
| SciCode (AA) | 66.9% | – |
| τ²-Bench Telecom (AA) | – | 96.5% |
| AIME (Vals) | – | 96.5% |
| LiveCodeBench | – | 84.3% |
| MMLU-Pro | – | 86.3% |
| LegalBench | – | 77.7% |
| CorpFin | – | 63.7% |
| TaxEval | – | 74.1% |
| MedQA | – | 94.5% |
| Terminal-Bench 2.1 (Vals) | – | 44.2% |
| SWE-bench (Vals) | – | 72.2% |
| GPQA Diamond (Vals) | – | 88.6% |
| Vals Index | – | 17.6 |
| Kagi LLM Benchmark | – | 75.0% |
| ARC-AGI-1 | – | 89.5% |
| ARC-AGI-2 | – | 65.1% |
| ARC-AGI-3 | – | 0.1% |
| CritPt | 31.7% | 6.6% |
| GDPval (AA) | 67.3% | – |
| APEX-Agents (AA) | – | 14.2% |
| CaseLaw v2 | – | 54.5% |
| Code Migration | – | 0.3% |
| Excel Modeling Benchmark | – | 11.7% |
| Finance Agent v2 | – | 28.5% |
| Harvey's Legal Agent Benchmark | – | 0.0% |
| Legal Research Bench | – | 13.9% |
| MedCode | – | 32.2% |
| MedScribe | – | 63.4% |
| MMMU-Pro (Vals) | – | 83.5% |
| MortgageTax | – | 45.4% |
| SAGE | – | 38.2% |
| Terminal-Bench 2.0 (Vals) | – | 40.5% |
| Vals Multimodal Index | – | 39.1% |
| Vibe Code Bench v1.1 | – | 4.1% |
| LMArena Maths | – | 1467 |
| LMArena Creative Writing | – | 1446 |
| LMArena Instruction Following | – | 1445 |
| LMArena Multi-turn | – | 1480 |
| LMArena Longer Queries | – | 1465 |
| LMArena Document | – | 1439 |
| Chess Puzzles | – | 24.0% |
| ForecastBench | – | 60.7% |
| Terminal-Bench 4.0 (AA) | 59.6% | – |
| AutomationBench | 69.5% | – |
| GDP.pdf | 26.2% | – |
| Harvey LAB | 91.2% | – |
| AA-Omniscience: accuracy | 66.2% | 28.9% |
| AA-Omniscience: non-hallucination | 41.4% | 82.6% |
| AA-Briefcase | 1822 | – |
Data as of 2026-09-22. Best configuration of each model; every score links to its source on the model pages.
Claude Opus 5.5 vs Grok 4.20: questions
- Is Claude Opus 5.5 better than Grok 4.20?
- Claude Opus 5.5 (thinking) leads on quality: 70.7 vs 59.7. The BenchLeader Index combines every independent quality benchmark; Claude Opus 5.5 (thinking) is ahead overall as of 2026-09-22, but check the category scores for your use.
- Which is cheaper, Claude Opus 5.5 or Grok 4.20?
- Grok 4.20 is cheaper: $1.56 against $8.00 per million tokens, blended at three input tokens per output token.
- Which is faster, Claude Opus 5.5 or Grok 4.20?
- Grok 4.20 streams faster: 103 against 48 output tokens per second.
- Which has the larger context window?
- Grok 4.20 accepts more context: 2M against 1M tokens.