GPT-4.1 mini

OpenAI launched GPT-4.1 mini in the API on April 14, 2025, alongside GPT-4.1 and GPT-4.1 nano, as part of a new series targeting coding, instruction following, and long-context tasks. The mini variant matches or exceeds GPT-4o in intelligence evaluations while cutting cost by 83% and nearly halving latency, making it positioned for low-latency interactive applications. All three models share a 1 million token context window and a refreshed June 2024 knowledge cutoff. GPT-4.1 mini carries the family's coding and instruction-following gains, with broader GPT-4.1 family benchmarks including 54.6% on SWE-bench Verified, 38.3% on MultiChallenge, and 72.0% on Video-MME long-no-subtitles. The series was shaped with developer input to optimize real-world utility, delivering stronger performance across the latency curve. The launch reflects OpenAI's broader push to expand cost-effective API options for production workloads.

Microsoft announced the GPT-4.1 model series, including 4.1-mini and 4.1-nano, on Azure AI Foundry and GitHub Models in an April 14, 2025 Azure blog post by VP of Product Management Steve Sweetman. The series emphasizes improved coding, instruction following, and long-context processing, with fine-tuning support enabled for GPT-4.1 and 4.1-mini so developers can customize them for specific business needs. The post details that GPT-4.1 supports one million token inputs for detailed, multi-step agent workflows, while retaining GPT-4o-family API features such as tool calling and structured outputs. GPT-4.1-mini is positioned as the balanced option between 4.1's higher cost and 4.1-nano's lowest cost, all sharing the same 1M context length for Azure OpenAI Service customers.

OpenAI's GPT-4.1-mini became generally available in GitHub Models alongside the GPT-4.1 launch, as a lightweight variant built for high performance at lower cost and latency. According to the GitHub Changelog announcement, it inherits GPT-4.1 series strengths including enhanced coding, improved instruction following, long-context understanding, and multimodal text-and-image support, making it suited to real-time and parallel or chained workloads. The same announcement notes that GPT-4.1-mini supports parallel function calling and structured output generation, positioning it for chatbots, coding copilots, and AI-powered agents in the GitHub Models playground and API. It is paired with an even lighter GPT-4.1-nano variant, giving GitHub developers a tiered choice for cost-sensitive, high-frequency tasks without losing the core GPT-4.1 capability set.

BenchLM's model catalog page for GPT-4.1 mini (data as of September 1, 2026) reports the model was released on April 14, 2025, is a proprietary non-reasoning model with a 1M token context window, and lists API pricing at $0.40 per million input tokens and $1.60 per million output tokens. The page assigns GPT-4.1 mini a The benchmark snapshot shows GPT-4.1 mini ranking 113th of 147 in Coding (23rd percentile), with only five published benchmark rows available and several tracked slots empty, which the page flags as a reason to validate before choosing the model for production. Other capability axes such as Agentic, Reasoning, Knowledg

OpenAI announced on January 29, 2026 that GPT-4o, GPT-4.1, GPT-4.1 mini, and OpenAI o4-mini will be retired from ChatGPT on February 13, 2026, alongside the previously announced retirement of GPT-5 (Instant and Thinking). The announcement explicitly names GPT-4.1 mini by its full version identifier and groups it with o Because the retirement is scoped to ChatGPT and explicitly preserves API availability, the direct lifecycle impact on the underlying GPT-4.1 mini model served via API endpoints is limited per the announcement's own language. The post also documents how user feedback around GPT-4o helped shape GPT-5.1 and GPT-5.2, with