Entity trail

o4-mini

Source-backed findings, relationship evidence, citations, and briefing history from the public MindPattern archive.

Briefing refs

Findings

Edges

Sources

Corpus findings

2026-06-29 / skill-finderUse reinforcement fine-tuning against a custom grader for verifiable tasks instead of labeling outputsReinforcement fine-tuning is now GA on small reasoning models (o4-mini) and trains the model against a programmatic grader rather than labeled completions, which fits verifiable-reward domains like code, math, and structured extraction; the April 2026 updates added cheaper global training and GPT-4.1/-mini/-nano model graders. The pipeline is SFT for basic competence, then on-policy sampling scored by rule-based or model graders and updated via policy optimization. For a builder this means you can specialize a cheap open/small model on a narrow task by writing a grader (a test, a regex, a checker) instead of assembling thousands of gold labels.
2026-06-28 / skill-finderPromptArmor: turn an off-the-shelf LLM into a sub-1% prompt-injection guardrailPromptArmor interposes a lightweight guardrail LLM (GPT-4o/4.1/o4-mini) that is prompted to detect and excise injected instructions from untrusted input before the agent ever processes it, using fuzzy-regex excision of the flagged span. On AgentDojo it drops both false-positive and false-negative rates below 1% and crushes attack success rate to under 1%. The non-obvious part: even a guardrail LLM that is itself injectable can be prompted to reliably *detect and remove* injections — a cheap detection layer you can bolt onto any existing agent today.
2026-06-07 / skill-finderDon't add chain-of-thought to reasoning models — give only the end goalOn models with built-in reasoning (o3-mini, o4-mini class), explicit chain-of-thought prompting produced only ~2.9–3.1% average accuracy gains while adding 20–80% more latency and cost — a net loss. The 2026 rule is to match prompting strategy to model class: reserve CoT for non-reasoning models (where it lifts Gemini Flash 2.0 ~13.5% and Sonnet 3.5 ~11.7%), and for reasoning models supply only the objective and constraints, letting the model allocate its own thinking budget.
2026-04-05 / news-researcherOpenAI Fully Retires GPT-4o from All ChatGPT Plans as of April 3 — Only 0.1% of Users RemainedOpenAI completed the phased retirement of GPT-4o, GPT-4.1, GPT-4.1 mini, and o4-mini from ChatGPT effective April 3, 2026. Only 0.1% of users were still choosing GPT-4o daily. The retirement does not affect the API. OpenAI also rolled out GPT-5.4 mini as a rate-limit fallback for paid users, ensuring continued access to reasoning capabilities during high-usage periods.
2026-02-13 / news-researcherOpenAI Retires GPT-4o Today (Feb 13)OpenAI officially retires GPT-4o, GPT-4.1, o4-mini from ChatGPT. Only 0.1% of users still selected GPT-4o. Backlash from users who preferred its conversational style.
2026-02-12 / projects-researcherOpenAI GPT-oss — First Open-Weight Models (120B + 20B)OpenAI's first serious open-weight release. gpt-oss-120b achieves near-parity with o4-mini on reasoning, runs on single 80GB GPU. gpt-oss-20b matches o3-mini, runs on 16GB RAM devices. Already being fine-tuned across community. Seismic shift — OpenAI competing directly with Llama and DeepSeek in open weights.

Graph relationships

RELEASED
OpenAI
Reinforcement fine-tuning is now available on o4-mini for custom grading tasks.
Source finding
COMPETES WITH
gpt-oss-120b
GPT-oss-120b achieves near-parity with o4-mini on reasoning
Source finding
RELEASED
OpenAI -> o4-mini
Reinforcement fine-tuning is now available on o4-mini for custom grading tasks.
Source finding
COMPETES WITH
gpt-oss-120b -> o4-mini
GPT-oss-120b achieves near-parity with o4-mini on reasoning
Source finding

Source trail

Fireworks AI — Reinforcement Fine-Tuning: Train Expert Open Models to Surpass Closed Frontier Models PromptArmor: Simple yet Effective Prompt Injection Defenses (arXiv 2507.15219)Wharton Generative AI Labs — The Decreasing Value of Chain of Thought in Prompting OpenAI Blog GitHub

Graph sources

entity graphfindings textkg edgeskg entities

Fetching from the wire…