GPT-5.4 mini
OpenAIProprietaryReasoningVerifiedReleased Mar 17, 2026Input: file, image, text
Overall score
53.4
Rank 82 of 222 · 42.6–64.2
Context window
400K
400,000 tokens
Price per 1M tokens
$0.75 in · $4.50 out
✓ Verified on OpenAI's pricing page
Blended price
$1.69
3:1 input:output tokens
Output speed
—
Not measured yet
Category scores
Scores per capability area with the model's rank among ranked models in that category.
- 1Instruction following#2789.9
- 2Multimodal#2873.8
- 3Knowledge#7760.9
- 4Coding#7859.4
- 5Agentic & tool use#8059.3
- 6Math55.9
- 7Reasoning43.7
Benchmark results (14)
As reported in system cards, papers and public evaluations. Position compares this model with every other model reporting the same benchmark.
Agentic & tool use
| MCP Atlas | 57.7 | #31 of 42 |
| Terminal-Bench 2.0 | 60.0 | #17 of 39 |
| OSWorld-Verified | 72.1 | #21 of 35 |
| Toolathlon | 42.9 | #16 of 22 |
Knowledge
| GPQA | 88.0 | #34 of 117 |
| HLE | 41.5 | #28 of 69 |
| HLE w/o tools | 28.2 | #34 of 43 |
Reasoning
| ARC-AGI-2 | 18.9 | #39 of 41 |
| ARC-AGI-1 | 63.70 | #35 of 36 |
Multimodal
| MMMU-Pro | 76.6 | #28 of 48 |
| MMMU-Pro w/ Python | 78.0 | #8 of 9 |
Math
| FrontierMath v2 (Tiers 1-3) | 51.228 | #41 of 64 |
| FrontierMath v2 (Tier 4) | 9.756 | #45 of 51 |
Coding
| FrontierCode 1.1 Main | 27.0 | #14 of 16 |
Pricing by source
USD per 1M tokens.
| Source | Input | Output | Cached in |
|---|---|---|---|
| ✓ Official (OpenAI) checked Oct 17, 2026 · gpt-5.4-mini | $0.75 | $4.50 | $0.075 |
| BenchLM list price | $0.75 | $4.50 | $0.075 |
| OpenRouter openai/gpt-5.4-mini | $0.75 | $4.50 | $0.075 |
Standard tier, prompts under 272K tokens.
OpenAI's current API pricing page lists gpt-5.4-mini at $0.75 input / $0.075 cached input / $4.50 output per million tokens.
Arena ratings
Human-preference ratings from LMArena. Best-rated variant shown.
| Text(high) | 1412 | #137 · 63.4K votes |
| WebDev (coding)(High) | 1397 | #99 · 11.9K votes |
| Vision(High) | 1245 | #69 · 25.7K votes |
Speed
Measured by us through OpenRouter (median of trials).
Not measured yet.
Related models
Same family
- GPT-5.466.7
- GPT-5.4 Pro68.7
- GPT-5.4 nano46.9
Similar score
- MAI-Thinking-153.3 · —
- GLM-5-Turbo53.2 · $1.90
- Grok 4.353.1 · $1.56
- Kimi K2.553.0 · $1.20
- Qwen3.7 Plus53.8 · $0.56