GPT-5.4 Mini
openai/gpt-5.4-mini- OpenAI
- Family · GPT-5 Mini
- Released 17 Mar 2026
- File · Image · Text → Text
- Reasoning · adjustable
- Tools · JSON · Vision
Spec & list price
Context391K
Max output128K
Input / 1M$0.75
Output / 1M$4.506× input
Scores · standing
Intelligence24.1#104/489
Coding56.1#53/196
Usage
Tokens / day7.7B▼ 22% 7d
Usage rank#98
Requests / day952.7K
Tool calls / req0.12
Per request · 30 days
Avg input6.8K0.5× median
Avg output2630.4× median
Cache hit58%0.8× median
Reasoning share24%0.4× median
Tool-call failures0.5%0.3× median
Consumption
Peak 61.5B 14 OctLow 5.3B 1 NovRank today #98
News & trends
5 mentions this week ▲ 400%this model · videos, posts, articles · 30 days
Trending · 14 days
- I built a PostGIS pipeline that finds the officials responsible for any spot in India, plus a GPT-5.4 mini agent that routes complaints to them126 upvotes
- My Laptop Model Tied GPT-5.4-mini and Was 9× Faster. Is It Cheaper? Self-Hosting vs API, Measured6 views
- I Fine-Tuned a 0.6B Model on My Laptop. It Matched GPT-5.4-mini on One Job: LoRA, Measured10 views
News · latest from OpenAI
- GPT-6.1 Terra vs. Claude Opus 5.5: Cheaper and faster in three tests at maxthenewstack.io · 2 Nov
- Chinese labs now target OpenAI's GPT-6 Astra for distillationfork.zone · 1 Nov
- GPT-6.1 Terrasimonwillison.net · 31 Oct
- GPT-5.6 Sol can beat Factorio on peaceful moderyanmadden.net · 31 Oct
Benchmarks
bar = share of the top score · Benchmark data: Artificial AnalysisSpeed · median across 5 providers, latest day
Output speed#78/33499 tok/s1.5× faster than median
Fastest provider117 tok/sbest of 5 providers
First token#1991.31 s1.3× longer than median
#26 cheapest of 34 with a similar score
Value
Intelligence Index
Blended price per 1M tokens (log)
Family progress
Intelligence Index
- Family releases24.1
Technical details
- Reasoning
- on · medium
- Tokenizer
- GPT
- Knowledge cutoff
- 2025-08-31
- Moderation
- moderated
Reasoning effort
- xhigh
- high
- medium · default
- low
- none
Request parameters 11
- tools
- tool_choice
- response_format
- structured_outputs
- reasoning
- reasoning_effort
- include_reasoning
- max_tokens
- max_completion_tokens
- seed
+1 more
- verbosity
Interesting facts
- 2× the output room of similar models (128.0K against a 65.5K median).
- Released 230 days ago, the newest of 2 GPT-5 Mini models.
When to use
Best for
- Cost-aware coding agents
- Tool-calling workflows with solid reasoning
- Long document analysis on a budget
- Image and file understanding
- Offline bulk jobs via the batch variant
Think twice for
- Hardest frontier reasoning
- New projects (usage is declining)
- Self-hosting (closed weights)
Research papers
- Component Translation and Judge-Labelled Refusal: A Multilingual Safety Evaluation28 Oct
- M4Bench: Evaluating Procedural Specification for Clinical EHR Derivation Agents27 Oct
- Validating Autoformalized Specifications via Contrapuntal Languages21 Oct
- Symbolic Constraint Agents: Mining Formal Tool-Call Constraints to Prevent Agent Errors18 Oct
- SkillNEO: Non-Elitist Evolutionary Optimization of Agent Skills14 Oct
- Smart Alone, Blind Together: Why Accuracy Fails to CertifyTheory of Mind in Multi-Agent Teams13 Oct
- Probable but Not Valid: A Psychometric Audit of Decision Models as Synthetic Survey Respondents10 Oct
- Veracity labels flip the emotional tone of language models but leave racial disparities intact8 Oct
Social
Most watched and discussed · 90 days
- I Fine-Tuned a 0.6B Model on My Laptop. It Matched GPT-5.4-mini on One Job: LoRA, MeasuredSlav Kurochkin · 10 views · 26 Oct
- My Laptop Model Tied GPT-5.4-mini and Was 9× Faster. Is It Cheaper? Self-Hosting vs API, MeasuredSlav Kurochkin · 6 views · 27 Oct
Same family
GPT-5 Mini- 20.6
Alternatives
other vendors · similar capability- 26.3
- 21.8
- T25.0
- 22.2
Using GPT-5.4 Mini at volume?
Sign in to ask for a volume discount. You will continue in the OSMEXO dashboard.