Anthropic launched Claude for Teachers on July 14, 2026, giving verified US K-12 educators free access to premium Claude capabilities, standards-mapped lesson planning through a Learning Commons connector, and both Claude Cowork and Claude Code. It is free through June 30, 2027, and it is Anthropic's direct answer to ChatGPT for Teachers, which OpenAI shipped eight months earlier.

That last part is the story nobody has actually reported. Every launch-day article, including Anthropic's own, describes what Claude for Teachers does. None of them puts it next to OpenAI's teacher plan line by line to show where the two actually diverge, and the gap matters more than the press release does.

Strip away the announcement language and Claude for Teachers is three things bundled together. First, free premium access to Claude for any US K-12 educator who verifies their status, available through applications closing June 30, 2027. Second, a connector to Learning Commons, which gives Claude the academic standards for all 50 states, plus the smaller learning competencies underneath each standard and the order students typically move through them. Third, a library of teaching skills co-developed with Learning Commons and refined with feedback from classroom teachers, including staff at Prospect Schools in Brooklyn.

The standards layer is the part worth pausing on. Most teachers who've asked a general-purpose chatbot for a lesson plan already know the output can look polished and still miss the actual scope-and-sequence their district follows. Anthropic's pitch is that alignment happens before the teacher sees the draft, not after, because Claude is drawing on the standard and the fine-grained competencies beneath it rather than guessing at grade-level from the prompt alone.

Anthropic and OpenAI each announced a free teacher plan, eight months apart, and both companies wrote long blog posts about their own product. Neither wrote about the other's. Pulled directly from each company's own announcement and help documentation, here is where the two plans actually differ:

Two things jump out once it's laid side by side. Claude leans on curriculum-standards depth as its wedge; ChatGPT leans on raw model capacity and file/collaboration breadth. Neither company's page mentions the other's approach, which is exactly why a teacher comparing the two on marketing copy alone would miss the actual trade-off: standards-first scaffolding versus a bigger general-purpose toolkit.

Buried past the privacy section of Anthropic's own announcement is the detail that separates Claude for Teachers from a chatbot with an education skin. Because the plan includes Claude Cowork and Claude Code, a teacher can hand Claude a folder of roster data, diagnostic scores, attendance, and personal notes, and have it build a picture of where every student stands, not as a one-off chat but as a task it can repeat.

Anthropic's own example: a teacher can schedule a recurring job that reviews the day's exit tickets every afternoon at 4pm and adjusts tomorrow's plan to match what students actually showed they understood, running automatically while the teacher drives home. That is a scheduled agent task, not a prompt-and-response exchange, and it is not something ChatGPT for Teachers' current feature set replicates; OpenAI's plan gives teachers Deep Research and unlimited GPT-5.1 access, but nothing framed as a standing, self-triggering classroom job.

The trade-off is real, though. A tool that reads roster and diagnostic data unattended raises the exact student-privacy question critics have already been asking about AI in classrooms. Anthropic's answer is that teachers control what gets shared and that nothing shared is used to train the model, backed by a K-12-specific Data Processing Addendum written to comply with FERPA.

Anthropic is piloting Claude for Teachers in Detroit Public Schools Community District, starting next school year. Drew Bent, Anthropic's education lead, said Detroit was already a "human-centric" user of other Claude products, which is why it got picked for the study on educator well-being and practice rather than just classroom output. Anthropic will train teachers at a handful of Detroit schools directly, then evaluate how the product changes day-to-day teaching before any wider K-12 rollout is even discussed.

One number worth sitting with: teachers who use AI weekly report saving close to six hours a week, according to Gallup survey data OpenAI cited in its own teacher-plan launch. Neither company has published a comparable number specific to standards-aligned planning versus general chat use, which is the actual gap a district would need filled before trusting either tool with a full week's prep load.

Claude for Teachers bets on curriculum depth. ChatGPT for Teachers bets on model breadth. Pick based on which gap costs your school more time.

Chalkbeat's reporting on the launch notes that AFT is working with OpenAI, Microsoft, and Anthropic on shared privacy standards, but notably not with Google, which has built its own separate AI training track for teachers and struck a statewide deal to bring Gemini into every K-12 school in Utah. That leaves three separate trust frameworks for a district to evaluate, not two, if Google is also on the shortlist.

Randi Weingarten, president of the American Federation of Teachers, is quoted in Anthropic's own release calling Claude for Teachers "a tool designed by and for educators to assist them instructionally." What the press release does not mention, and what Chalkbeat's coverage does note in passing without connecting it back to this launch, is that Weingarten spent the prior weeks calling for a full ban on student-facing AI in elementary grades, then toured a Newark school to see a rival product, Khan Academy's Khanmigo, in action, a day before making that call.

Put those two data points together and the AFT's actual position is more specific than "pro-AI" or "anti-AI": union leadership is comfortable with AI as a teacher's tool and uneasy with AI as a student's tool, at least at the elementary level. That distinction is exactly the line Anthropic drew when it built Claude for Teachers as a teacher-only product; Anthropic has confirmed most K-12 students cannot access Claude at all, due to the platform's 18-and-over policy. The union's stance and the product's design boundary line up almost exactly, which is worth noting given that AFT is also working with Anthropic and OpenAI, but not Google, on shared privacy standards.

If your school already has a ChatGPT for Teachers workspace, wholesale switching for one feature is rarely worth the disruption. The honest read: Claude for Teachers is the stronger pick if your prep time is mostly lost to checking standards alignment by hand, since that is the one feature ChatGPT's plan does not attempt to replicate. ChatGPT for Teachers is the stronger pick if your school already runs on Microsoft 365 or Google Workspace and values raw model capacity, unlimited GPT-5.1 access, and Deep Research over curriculum-mapped scaffolding.

Both are free through roughly the same window, both claim FERPA-aligned privacy postures with no default training on educator data, and both are explicitly built for teachers, not students. Neither is a reason to abandon the other if it is already working; they solve adjacent but different problems.

Yes. It is free for verified US K-12 educators, with applications open through June 30, 2027.

Verified K-12 educators in the United States only. Consistent with Claude's 18-and-over policy, most students cannot access it directly.

No. Anthropic states that Claude for Teachers data is not used for model training, and student information is covered by a K-12 Data Processing Addendum written to comply with FERPA.

Claude for Education is the university-wide plan for higher-ed institutions like Northeastern, LSE, and Champlain College. Claude for Teachers is the new K-12-specific plan for individual verified teachers, announced July 14, 2026.

Yes, along with Claude Cowork, both usable for tasks like scheduled data review and multi-file lesson generation, not just single-turn chat.

This comparison draws on Anthropic's own Claude for Teachers announcement, OpenAI's ChatGPT for Teachers announcement, and Chalkbeat's reporting on the launch. Anthropic's K-12 privacy standards are detailed separately on its site if you want the DPA language before rolling this out district-wide.