Grok 4.3 vs o3
Grok 4.3 
o3 Verdict
- Grok 4.3 (medium) and o3 are level on quality (60.8 vs 61.6).
- Grok 4.3 (medium) is stronger in composite, instruction following, knowledge, long context, multimodal.
- o3 is stronger in agents & tools, coding, human preference, maths, reasoning.
- Grok 4.3 (medium) is 2.2× cheaper ($1.56 vs $3.50 per 1M blended).
| Metric | Grok 4.3 (medium) | o3 |
|---|---|---|
| BenchLeader Index | 60.8 | 61.6 |
| Agents & tools score | 60.1 | 70.3 |
| Composite score | 60.3 | 54.5 |
| Instruction following score | 80.0 | 65.0 |
| Knowledge score | 72.2 | 58.0 |
| Long context score | 64.0 | 63.8 |
| Multimodal score | 60.3 | 57.9 |
| Coding score | – | 62.0 |
| Human preference score | – | 62.3 |
| Maths score | – | 78.6 |
| Reasoning score | – | 61.5 |
| Blended price $/M | $1.56 | $3.50 |
| Output speed | 121 tok/s | 107 tok/s |
| Time to first answer | 11.7 s | 6.3 s |
| Context window | 1M | 200k |
| Epoch Capabilities Index | – | 146.9 |
| LMArena Text | – | 1432 |
| LMArena Hard Prompts | – | 1441 |
| LMArena Coding | – | 1460 |
| LMArena Vision | – | 1214 |
| AA Intelligence Index | 24.8 | 20.2 |
| IFBench | 83.3% | 71.4% |
| AA-LCR | 75.0% | 74.7% |
| MMMU-Pro | 75.8% | 70.1% |
| AA-Omniscience | 16.7 | -15.6 |
| Terminal-Bench Hard | 30.3% | 37.1% |
| GPQA Diamond (AA) | 89.0% | 82.7% |
| Humanity's Last Exam (AA) | 30.0% | 20.1% |
| τ²-Bench Telecom (AA) | 91.2% | 80.7% |
| PRBench Finance | – | 47.7% |
| PRBench Legal | – | 48.6% |
| MMMU (validation) | – | 82.9% |
| MMMU-Pro (official) | – | 76.4% |
| Kagi LLM Benchmark | – | 67.6% |
| IFEval (HELM) | – | 86.9% |
| Omni-MATH (HELM) | – | 71.4% |
| WildBench (HELM) | – | 86.1% |
| MMLU-Pro (HELM) | – | 85.9% |
| GPQA Diamond (HELM) | – | 75.3% |
| HELM Capabilities mean | – | 81.1% |
| Aider Polyglot | – | 81.3% |
| SWE-bench Verified (bash only) | – | 58.4% |
| SWE-bench Verified (any scaffold) | – | 58.4% |
| BFCL Overall | – | 63.0% |
Data as of 2026-09-12. Best configuration of each model; every score links to its source on the model pages.
Grok 4.3 vs o3: questions
- Is Grok 4.3 better than o3?
- Grok 4.3 (medium) and o3 are level on quality (60.8 vs 61.6). The BenchLeader Index combines every independent quality benchmark; o3 is ahead overall as of 2026-09-12, but check the category scores for your use.
- Is Grok 4.3 better than o3 for agentic tasks?
- o3 scores higher in agentic tasks (70 vs 60 on the category index, where 50 is average).
- Which is cheaper, Grok 4.3 or o3?
- Grok 4.3 is cheaper: $1.56 against $3.50 per million tokens, blended at three input tokens per output token.
- Which is faster, Grok 4.3 or o3?
- Grok 4.3 streams faster: 121 against 107 output tokens per second.
- Which has the larger context window?
- Grok 4.3 accepts more context: 1M against 200k tokens.