GPT-5.4 nano
| Released | 2026-03-17 |
|---|---|
| Price per 1M | $0.20 in · $1.25 out |
| Overall score | 50.1 · rank 113 |
| Context window | 400,000 tokens |
| Provider | OpenAI |
| Weights | Proprietary |
Scores
| Dimension | Score | Rank | Benchmarks |
|---|---|---|---|
| Overall: Balanced | 50.1 | 113 | 24 |
| Overall: Coding | 51.7 | 98 | 13 |
| Overall: Reasoning | 49.8 | 116 | 14 |
| Overall: Agentic work | 51.2 | 106 | 12 |
| Overall: Frontend | 49.1 | 98 | 6 |
| Coding | 53.6 | 90 | 3 |
| Agents | 50.2 | 106 | 2 |
| Reasoning | 51.2 | 108 | 7 |
| Math | 54.0 | 80 | 4 |
| Knowledge | 39.2 | 142 | 3 |
| Frontend | 47.2 | 81 | 1 |
| General | 49.1 | 117 | 2 |
| Multimodal | 47.3 | 86 | 2 |
| Value vs price | +4.6 | 80 |
Configurations
| Configuration | Balanced | Benchmarks |
|---|---|---|
| Unspecified setting | 43.6 ±7.0 | 2 |
| High | 49.8 ±1.2 | 20 |
| Low | 47.3 ±4.9 | 7 |
| Medium | 45.9 ±8.0 | 3 |
| Non-reasoning | 41.0 ±5.9 | 5 |
| Extra High | 53.1 ±4.9 | 5 |
Evidence
Coding 53.6 ±2.9
| Benchmark | Result | Source | Date |
|---|---|---|---|
| SciCode | 47.2% | Epoch AI | 2026-11-11 |
| SWE-bench Verified (Vals run) | 69.8% | Vals AI | 2026-09-01 |
| LiveCodeBench (Vals run) | 84.0% | Vals AI | 2026-09-01 |
Agents 50.2 ±2.6
| Benchmark | Result | Source | Date |
|---|---|---|---|
| Terminal-Bench 2.1 (Vals run) | 41.6% | Vals AI | 2026-10-01 |
| Tax Agent Bench | 5.0% | Vals AI | 2026-11-09 |
Reasoning 51.2 ±2.9
| Benchmark | Result | Source | Date |
|---|---|---|---|
| ARC-AGI-2 | 5.7% | Epoch AI | 2026-11-11 |
| ARC-AGI-2 (ARC Prize verified) | 5.7% | ARC Prize | 2026-11-11 |
| CritPt | 9.3% | Epoch AI | 2026-11-11 |
| ARC-AGI-1 (ARC Prize verified) | 51.5% | ARC Prize | 2026-11-11 |
| ARC-AGI-1 | 51.5% | Epoch AI | 2026-11-11 |
| GPQA Diamond | 78.5% | Epoch AI | 2026-04-14 |
| GPQA Diamond (Vals run) | 77.5% | Vals AI | 2026-09-01 |
| Kagi Reasoning Benchmark | 39.7% | Kagi | 2026-11-11 |
| Chess Puzzles | 45.0% | Epoch AI | 2026-10-14 |
| Mystery Game Puzzles | 9.0% | Epoch AI | 2026-08-27 |
Math 54.0 ±1.4
| Benchmark | Result | Source | Date |
|---|---|---|---|
| FrontierMath Tier 4 | 12.2% | Epoch AI | 2026-06-12 |
| FrontierMath Tiers 1-3 | 44.9% | Epoch AI | 2026-06-12 |
| ProofBench | 5.0% | Epoch AI | 2026-11-11 |
| OTIS Mock AIME 2024-2025 | 87.8% | Epoch AI | 2026-04-14 |
| FrontierMath Tiers 1-3 (2025 set) | 25.9% | Epoch AI | 2026-04-15 |
| FrontierMath Tier 4 (2025 set) | 6.3% | Epoch AI | 2026-04-15 |
Knowledge 39.2 ±3.6
| Benchmark | Result | Source | Date |
|---|---|---|---|
| SimpleQA Verified | 11.7% | Epoch AI | 2026-08-27 |
| MMLU-Pro (Vals run) | 77.2% | Vals AI | 2026-09-01 |
| Legal Research Bench | 6.3% | Vals AI | 2026-11-09 |
Frontend 47.2 ±0.8
| Benchmark | Result | Source | Date |
|---|---|---|---|
| Vibe Code Bench v1.1 | 26.1% | Vals AI | 2026-11-09 |
General 49.1 ±3.3
| Benchmark | Result | Source | Date |
|---|---|---|---|
| LMArena Text | 1,401 | LMArena | 2026-11-09 |
| EQ-Bench Longform Creative Writing | 57.8% | EQ-Bench | 2026-11-11 |
Multimodal 47.3 ±2.3
| Benchmark | Result | Source | Date |
|---|---|---|---|
| MMMU-Pro (Vals run) | 73.6% | Vals AI | 2026-09-01 |
| LMArena Vision | 1,199 | LMArena | 2026-11-09 |
Results not used in the score
| Benchmark | Result | Source | Why not used |
|---|---|---|---|
| ARC-AGI-1 | 18.3% | Epoch AI | A stronger configuration result was used |
| ARC-AGI-1 | 38.2% | Epoch AI | A stronger configuration result was used |
| ARC-AGI-1 | 33.0% | Epoch AI | A stronger configuration result was used |
| ARC-AGI-1 (ARC Prize verified) | 18.3% | ARC Prize | A stronger configuration result was used |
| ARC-AGI-1 (ARC Prize verified) | 38.2% | ARC Prize | A stronger configuration result was used |
| ARC-AGI-1 (ARC Prize verified) | 33.0% | ARC Prize | A stronger configuration result was used |
| ARC-AGI-2 | 1.5% | Epoch AI | A stronger configuration result was used |
| ARC-AGI-2 | 3.6% | Epoch AI | A stronger configuration result was used |
| ARC-AGI-2 | 1.9% | Epoch AI | A stronger configuration result was used |
| ARC-AGI-2 (ARC Prize verified) | 1.5% | ARC Prize | A stronger configuration result was used |
| ARC-AGI-2 (ARC Prize verified) | 3.6% | ARC Prize | A stronger configuration result was used |
| ARC-AGI-2 (ARC Prize verified) | 1.9% | ARC Prize | A stronger configuration result was used |
| Chess Puzzles | 2.0% | Epoch AI | A stronger configuration result was used |
| Chess Puzzles | 20.0% | Epoch AI | A stronger configuration result was used |
| Chess Puzzles | 36.0% | Epoch AI | A stronger configuration result was used |
| FrontierMath Tiers 1-3 | 4.6% | Epoch AI | A stronger configuration result was used |
| FrontierMath Tiers 1-3 | 20.4% | Epoch AI | A stronger configuration result was used |
| GPQA Diamond | 55.6% | Epoch AI | A stronger configuration result was used |
| GPQA Diamond | 72.2% | Epoch AI | A stronger configuration result was used |
| Mystery Game Puzzles | 3.0% | Epoch AI | A stronger configuration result was used |
| Mystery Game Puzzles | 5.0% | Epoch AI | A stronger configuration result was used |
| Mystery Game Puzzles | 6.0% | Epoch AI | A stronger configuration result was used |
| OTIS Mock AIME 2024-2025 | 46.7% | Epoch AI | A stronger configuration result was used |
| OTIS Mock AIME 2024-2025 | 68.9% | Epoch AI | A stronger configuration result was used |
Names matched to this model
gpt-5-4-nano, gpt-5-4-nano-2026-03-17