GPT-5.4 mini scores 31.2 on the Respan Index (#63), ahead of Inkling at 30.1 (#64). Inkling leads on 9 of the 18 benchmarks both report. Both cost about the same per token at list price. Inkling has the larger context window (1M tokens).
* Estimated: no published scores in that category. How the Respan Index works
| 41.7% |
| 49.4% |
| LiveBench Coding (LiveBench) | 71.6% | 71% |
| LiveBench Data Analysis (LiveBench) | 70.8% | 72.8% |
| LiveBench Instruction Following (LiveBench) | 59.8% | 70.1% |
| LiveBench Language (LiveBench) | 71% | 73.5% |
| LiveBench Mathematics (LiveBench) | 78.5% | 88.4% |
| LiveBench Reasoning (LiveBench) | 71.3% | 78.3% |
| OTIS Mock AIME 2024-2025 (Epoch AI) | 88.9% | 88.9% |
| SAGE (Vals AI) | 50.8% | 36.6% |
| SimpleQA Verified | 29.4% | 43.9% |
| Terminal-Bench 4.0 (Vals AI) | 2.5% | 0.5% |
| WeirdML (Håvard Tveit Ihle) | 60.3% | 32.3% |
| LMArena Elo | 1447 | 1441 |
| LMArena WebDev (LMArena) | 1397 | 1413 |
Higher is better. Each score links to where it was published.