Grok 4.3 vs Grok 4.5
Respan Index by category
Each category is scored 0 to 100, and how far ahead the winner is. Leads under a point are close to even.
| Category | Grok 4.3 | Grok 4.5 |
|---|---|---|
| Coding | 5.2 Grok 4.3 | 25.5 Grok 4.5 |
| Agents | 4.4 Grok 4.3 | 64.7 Grok 4.5 |
| Reasoning | 36.2 Grok 4.3 | 42.8 Grok 4.5 |
| Knowledge & language | 33.0 Grok 4.3 | 49.7 Grok 4.5 |
| Human preference | 50.0 Grok 4.3 | 55.3 Grok 4.5 |
Benchmarks both report
Grok 4.5 leads on 17 of the 19 benchmarks both report. Higher is better. Each score links to where it was published.
| Benchmark | Grok 4.3 | Grok 4.5 |
|---|---|---|
| LMArena Elo | 1442 | 1466 |
| LMArena WebDev (LMArena) | 1357 | 1552 |
| LMArena Vision (LMArena) | 1242 | 1280 |
| Chess Puzzles (Epoch AI) | 18% | 35% |
| FrontierMath Tier 4 v2 (Epoch AI) | 14.6% | 24.4% |
| FrontierMath Tiers 1-3 v2 (Epoch AI) | 42.8% | 57.2% |
| GPQA Diamond | 88.8% | 93.4% |
| LiveBench Agentic Coding (LiveBench) | 18.5% | 56.5% |
| LiveBench Coding (LiveBench) | 69.9% | 68.6% |
| LiveBench Data Analysis (LiveBench) | 55.8% | 73% |
| LiveBench Instruction Following (LiveBench) | 62.8% | 71.5% |
| LiveBench Language (LiveBench) | 73.6% | 82.8% |
| LiveBench Mathematics (LiveBench) | 84.3% | 90.8% |
| LiveBench Reasoning (LiveBench) | 70.8% | 87.2% |
| OTIS Mock AIME 2024-2025 (Epoch AI) | 93.3% | 97.8% |
| SAGE (Vals AI) | 19.7% | 35% |
| SimpleQA Verified | 33.2% | 48.3% |
| Vending-Bench 2 (Andon Labs) | 35.26 | 3887.43 |
| WeirdML (Håvard Tveit Ihle) | 49.9% | 46.4% |
FAQ
Is Grok 4.3 better than Grok 4.5?
Grok 4.5 scores 46.5 on the Respan Index (#51), ahead of Grok 4.3 at 22.3 (#115). Grok 4.5 leads on 17 of the 19 benchmarks both report.
Which is cheaper, Grok 4.3 or Grok 4.5?
Grok 4.3 lists at $1.25 input and $2.5 output per 1M tokens; Grok 4.5 at $2 and $6. Grok 4.3 is 1.9x cheaper per token at list price.
Which has the larger context window?
Grok 4.3 accepts 1M tokens and Grok 4.5 500K tokens.