GPT-5.4 mini Benchmark Details
GPT-5.4 mini currently shows benchmark results led by Terminal Bench Hard (19 / 244, score 52.30), IF Bench (44 / 282, score 73.30), SAGE (10 / 64, score 50.81). This page also compares it with 2 competitor models and 1 predecessor or same-series models, including performance and pricing views when available.
Benchmark Results
Benchmark Results
Abstract Generalization
8 evaluationsKnowledge Exams
5 evaluationsScientific Reasoning
5 evaluationsMathematics
5 evaluationsRepository Engineering
1 evaluationsService Workflows
5 evaluationsInstruction Following
3 evaluationsCross-capability Suites
5 evaluationsAgentic Development
6 evaluationsTool Orchestration
4 evaluationsMemory & Persistence
5 evaluationsLong Reasoning
3 evaluationsCapability Indices
2 evaluationsCross-industry Work
2 evaluationsVisual Understanding
3 evaluationsLegal
3 evaluationsFinance
2 evaluationsCode Generation & Editing
1 evaluationsCompetitor Comparison
Benchmark scores for GPT-5.4 mini compared against top models in its class
12 benchmarks with comparable scores. Each model shows its best score; mode label is displayed below.
| Benchmark | GPT-5.4 miniCurrent | Haiku 4.5 | Gemini 3.0 Flash |
|---|---|---|---|
63.67Thinking Level · Extra High ｜ Tools | 47.67Extended Thinking | -- | |
18.90Thinking Level · Extra High ｜ Tools | 4.50Extended Thinking | 33.60Thinking Enabled | |
41.50Thinking Level · Extra High ｜ Tools | 10.40Thinking Enabled | 43.50Thinking Enabled ｜ Tools | |
10.00Thinking Level · Extra High | -- | 8.60Thinking Enabled | |
87.50Thinking Level · Extra High | 73.30Extended Thinking | 90.40Thinking Enabled | |
2.10Thinking Level · High | 2.1032K | 4.20Standard Mode | |
54.40Thinking Level · Extra High ｜ Tools | 39.45Extended Thinking ｜ Tools | 49.60Thinking Level · High ｜ Tools | |
50.81Thinking Level · Extra High ｜ Tools | 31.82Thinking Enabled ｜ Tools | -- | |
83.30Thinking Level · Extra High ｜ Tools | 54.70Thinking Enabled ｜ Tools | 91.23Thinking Level · High ｜ Tools | |
25.60Thinking Level · Extra High ｜ Tools | 9.30Thinking Enabled ｜ Tools | 27.32Thinking Level · High ｜ Tools | |
73.30Thinking Level · Extra High | 54.30Thinking Enabled | 78.00Thinking Enabled | |
66.37Deep Thinking Mode | 61.3264K | 72.40Thinking Level · High |
Standard API Pricing: GPT-5.4 mini vs. Peer Models
Shows standard text input and output pricing side by side for each model. If extended-context pricing exists, the chart keeps the base rate and explains the threshold below.
Source: DataLearnerAI. Standard text prices shown here use the default supplier. · USD / 1M tokens
| Model | Supplier | Standard input | Standard output | Base price applies to |
|---|---|---|---|---|
GPT-5.4 mini | OpenAI | $0.75 / 1M tokens | $4.5 / 1M tokens | — |
Haiku 4.5 | Anthropic | $1 / 1M tokens | $5 / 1M tokens | — |
Gemini 3.0 Flash | Google DeepMind | $0.5 / 1M tokens | $3 / 1M tokens | — |
Version History
How each version of the GPT-5.4 mini series stacks up on benchmark tests
12 benchmarks with comparable scores. Each model shows its best score; mode label is displayed below.· Click a row to view its trend chart.
| Benchmark | GPT-5.4 miniCurrent | GPT-5-mini |
|---|---|---|
63.67Thinking Level · Extra High ｜ Tools | 54.33Thinking Level · High ｜ Tools | |
18.90Thinking Level · Extra High ｜ Tools | 4.44Thinking Level · High ｜ Tools | |
41.50Thinking Level · Extra High ｜ Tools | 21.50Thinking Level · High | |
10.00Thinking Level · Extra High | 1.40Thinking Level · Medium | |
87.50Thinking Level · Extra High | 82.80Thinking Level · High | |
1661.80Standard Mode | 1310.30Standard Mode | |
2.10Thinking Level · High | 6.30Thinking Level · High | |
9.76Thinking Level · Extra High | 12.20Thinking Level · High | |
51.23Thinking Level · Extra High | 46.67Thinking Level · High | |
50.81Thinking Level · Extra High ｜ Tools | 42.99Thinking Level · High ｜ Tools | |
83.30Thinking Level · Extra High ｜ Tools | 71.10Thinking Level · Medium ｜ Tools | |
25.60Thinking Level · Extra High ｜ Tools | 15.50Thinking Level · High ｜ Tools |
Single-Benchmark Version Trend
Viewing: ARC-AGI-1 · Abstract Generalization
Standard API Pricing Across the GPT-5.4 mini Series
Shows standard text input and output pricing side by side for each model. If extended-context pricing exists, the chart keeps the base rate and explains the threshold below.
Source: DataLearnerAI. Standard text prices shown here use the default supplier. · USD / 1M tokens
| Model | Supplier | Standard input | Standard output | Base price applies to |
|---|---|---|---|---|
GPT-5.4 mini | OpenAI | $0.75 / 1M tokens | $4.5 / 1M tokens | — |
GPT-5-mini | OpenAI | $0.25 / 1M tokens | $2 / 1M tokens | — |