gpt-5.4-mini-2026-03-17| Provider | Input, $ per 1M | Output, $ per 1M | In our data since |
|---|---|---|---|
| 302.AI | $0.75 | $4.5 | 31 Jul 2026 |
| Microsoft Foundry | $0.75 | $4.5 | 31 Jul 2026 |
| OpenAI | $0.75 | $4.5 | 31 Jul 2026 |
Only the base tier and only per-token prices are shown. Batch, discounted and cached rates, as well as prices per image or per second of video, do not go into this table: they cannot stand in the same column as a price per million tokens.
The date in the last column is the day this price first entered our collection. The price may well be older: before that day we simply were not recording it. It has not changed since — otherwise a new row with a new date would stand in its place.
| Task set | Result | Run conditions | Measured by |
|---|---|---|---|
| Mock AIME 2024–2025 — olympiad problems | 87.21 % | effort: high · with a tuned harness | Epoch evaluations |
| GPQA Diamond — graduate-level questions | 78.11 % | effort: high · with a tuned harness | Epoch evaluations |
| ARC-AGI — generalising to unseen patterns | 63.67 % | effort: xhigh · with a tuned harness | https://arcprize.org/leaderboard |
| WeirdML — unusual machine learning tasks | 60.3 % | effort: high | https://htihle.github.io/weirdml.html |
| FrontierMath, levels 1–3 | 51.23 % | effort: xhigh | Epoch evaluations |
| DeepResearch Bench — deep research | 36.26 % | effort: low | https://drb.futuresearch.ai/#drb self-reported |
| SimpleQA Verified — factual accuracy | 28.6 % | effort: high | Epoch evaluations |
| APEX-Agents | 24.6 % | — | — |
| ARC-AGI-2 | 18.9 % | effort: xhigh | — |
| Chess puzzles | 13.72 % | effort: high | Epoch evaluations |
| CritPt — physics problems | 10 % | effort: xhigh | — |
| FrontierMath, level 4 — research-grade problems | 9.76 % | effort: xhigh | Epoch evaluations |