GPT-4.1
GPTBalanced
Tool UseVisionStructured Output
About this model
GPT-4.1 is a flagship large language model optimized for advanced instruction following, real-world software engineering, and long-context reasoning. It supports a 1 million token context window and outperforms GPT-4o and GPT-4.5 across coding (54.6% SWE-bench Verified), instruction compliance (87.4% IFEval), and multimodal understanding benchmarks. It is tuned for precise code diffs, agent reliability, and high recall in large document contexts, making it ideal for agents, IDE tooling, and enterprise knowledge retrieval.
Performance Tier
Balanced
GPT-4.1 is a balanced model from GPT : strong performance at a reasonable price.
Strong cost-performance ratio. Reliable for most professional use cases without premium pricing.
Pricing
This model is included in Elosia plans| Type | per 1M tokens |
|---|---|
| Input (prompt) | $2.00 |
| Output (completion) | $8.00 |
| Cache read | $0.500 |
Capabilities
Context Length1.0M
Max Output Tokens33K
TokenizerGPT
Inputimage, text, file
Outputtext
Release DateApril 14, 2025
Benchmarks
General Intelligence
MMLU
87.2%
Mathematics
MATH-500
78%
Programming
HumanEval
90.2%
SWE-bench Verified
54.6%
Reasoning
Humanity's Last Exam
5.4%
Recommended Use Cases
CodingGeneral ChatAnalysisData Extraction
Strengths
- Reliable general-purpose model with strong instruction following
- Large 1M token context window
- Good code generation abilities (HumanEval 90.2%)
- Competitive pricing for its capability level
Limitations
- Surpassed by GPT-5 family on most benchmarks
- No native reasoning/thinking mode
Resources
This model may use your data for training