When I shared @harvey’s model strategy a few months ago, there were two parts: 1. Build our own model 2. Use that to help customers do the same We’ve done the first. Now we’re hiring for the second: Harvey’s Private Model Program. We’re seeing huge demand from law firms to own their own intelligence leveraging private data. This will enable them to become frontier firms that get smarter with every client matter. You’ll:
 - Partner with law firms to build these systems. - Build the team that delivers this at scale.
 - Work with our technical org to define the platform that powers this team. We’re looking for a technical PM or founder type to own Harvey’s Private Model Program. DM me if this sounds interesting.

Harvey’s custom agents now run directly in Word. Run your agents on the document you already have open, review its suggestions, and apply edits as tracked changes - all without leaving Word.

Coming soon to Harvey: GPT-6 Sol. GPT-6 Sol showed particular strength on legal tasks focused on tax, immigration, and healthcare and life sciences, with additional gains in insurance and funds and asset management over 5.6 Sol. GPT-6 Luna will also be available soon in Harvey.

The hardest thing about building @harvey is doing what’s best for our customers despite immense pressure to do what’s easy. The easy thing would have been to force our customers onto consumption pricing before they were ready and serve them worse models to protect our margins. We chose to help our customers transition on a timeline that works for them and give them the best models in the meantime, even though it hurt our margins. This meant optimizing our product through routing, harness improvements, and post-training so we could serve frontier intelligence at an affordable price. It also meant building the infrastructure for customers to monitor and manage spend: usage dashboards, per-matter cost attribution, spend caps, and ROI reporting. As a result, we improved our gross margins from -50% to positive in a single quarter despite usage doubling month over month and continuing to serve the best models. Our philosophy is simple: do what’s best for our customers, even when it’s painful, hurts our margins, or draws criticism from competitors, X, and the press. We believe the most important part of building a company is earning and keeping your customers’ trust. You do that by doing the hard thing for them, even when it costs you.

Coming soon to Harvey: Claude Opus 5.5. In our evaluations, Opus 5.5 showed improved performance on agentic legal tasks, with gains over Opus 5 in data privacy and cybersecurity, emerging companies and venture capital, and capital markets.

Excited to try Jev in parts of our systems that need calibrated probabilities for categorical decisions. We currently use a hacky version of this idea: small LLM classifiers for routing, citations, parts of Vault, tool use, and user escalation. One challenge is that LLM softmax probabilities aren’t necessarily calibrated confidence estimates. It will be interesting to see how RLCD improves calibration over the naive approach. Jev doesn’t generate text, so its “hallucination-free” framing isn’t a full solution to hallucinations. But better routing, citation selection, and escalation could reduce hallucinations across the broader system. Longer term applications for law firms include matter selection, associate staffing, and predicting billing disputes. Also excited to see open-source implementation of RLCD so we can post-train these models ourselves.

Our cofounder @gabepereyra joined @brendanfoody to talk about: - Harvey's origin story - Becoming a full-stack AI company - Scaling data with @mercor - Helping law firms build and own their intelligence Full conversation:

The gold standard for evaluating complex legal work is partner review. Partners can cost over $3,000 an hour and associates $1,000, making expert review expensive at scale. Evaluating a model on LAB through expert review alone would cost millions of dollars. In practice, we combine rubric-based LLM scoring with sampled human preference. But rubrics miss errors beyond predefined criteria, and human reviewers get fatigued and make mistakes at scale. We built a generative reward model to bridge this gap by training agents to approximate partner review. We give these agents the original outputs, web search, and other tools to check our systems’ work. We find that their judgments correlate strongly with expert lawyer review. This approach will help us scale human review across model training and production products, and will be central to training Tenet 1.5.

Generative reward models are a promising direction for scaling up human judgment. Quality judgments for legal work product depend on many preference drivers like style, tone, writing quality, and document formatting that are less well-represented in current RLVR configurations vs. substantive and objective criteria. Legal outcomes are also inherently subjective — lawyers write documents called opinions for a reason. Trial judges and appellate judges disagree on 10-15% of cases. For this reason, our benchmarking + evals have historically relied on human lawyer preference judgments (SxS, Likert) which are high signal but low volume. In this article, @ItsJulioPereyra lays out how we’re scaling preference judgments with GRMs and their implications for both eval and training.

We built generative reward models (GRMs) to scale how human experts judge legal work product. A critical part of how we evaluate models and agent systems at Harvey is legal expert side-by-side review. Lawyers on our Applied Legal Research team and at our data partners compare two responses to the same task, choose which they prefer, and explain why. This process is difficult to scale: 1) Expert time is scarce and valuable, and 2) Lawyers often disagree on how to weigh the strengths and weaknesses of the model outputs. GRMs are LLM judges that compare two responses by generating a rubric for each task, and scoring each response against the rubric. We extended this approach with: 1) Sub-agents that let the GRM verify factual claims against the environment, source materials, and the web 2) Lawyer-defined rubric dimensions covering topics like accuracy, coverage, reasoning, style, and tone. GRMs help us scale expert judgment by serving as copilots: they review outputs, identify key differences, and highlight issues that warrant closer expert attention. In our early experiments, they largely reproduced lawyers’ model rankings. Deep dive by @ItsJulioPereyra on our early experiments with GRMs and how we're using agents to scale our data and eval efforts:

Thrilled to be joining Harvey and working in an area that's so integral to society. AI is a transformative technology, but realizing its potential requires going deep in specific domains and working in close partnership with the professionals who live and breathe that work. Legal is an especially rich and consequential place to do that. Harvey's progress, combined with the rapidly maturing ecosystem of open-weight models and post-training infrastructure, creates an enormous opportunity to push the frontier and make AI transformative in practice. We're building out Harvey's founding research team and hiring across post-training skill sets for people who want to work at the intersection of research and product. If that's you, please reach out.

Excited to welcome @asadovsky as Harvey’s Chief Research Officer. Before Harvey, Adam co-led post-training at Microsoft AI and Google DeepMind. As a CVP at Microsoft AI, he helped build MAI-Thinking-1, Microsoft’s reasoning model. As part of Gemini’s leadership team he helped train Gemini 1.0 through 2.5, including fine-tuning, RL, data, and evals. His prior work as a Distinguished Engineer at Google spanned Assistant, Search Quality, and Search Infrastructure. I met Adam three years ago when I sent him a cold LinkedIn DM and was surprised he responded. At a time when most dismissed the application layer and legal, Adam was curious and generous with his time. He quickly became someone I regularly turned to for advice on AI as we scaled Harvey over the past three years. When we first met, we were too early to hire someone of his caliber and scale, but I always hoped we’d eventually work together. As Winston and I got to know him better, what stood out even beyond his technical achievements was his character. Despite his incredible technical career, he remains curious, humble, practical, and cares deeply about the teams he builds. We couldn’t think of a better leader to help us build frontier intelligence for the professionals and institutions we serve.

We're excited to welcome @asadovsky as our Chief Research Officer. Prior to Harvey, Adam co-led post-training at Microsoft AI and Google DeepMind. As CVP at Microsoft AI, he led post-training of MAI-Thinking-1. As a Distinguished Engineer at Google DeepMind, he led teams working on Gemini 1.0 through Gemini 2.5. Earlier at Google, he led teams working on Assistant, Search Quality, and Search Infrastructure.

.@Harvey hired 1,000 people in 12 months, and will hire 600 more before EOY. 70% of the team joined in 2026. Their valuation's grown to $15.5B in four years. We asked VP Talent Maggie Landers how to build a company culture where 1,000 new hires can run as fast as possible:

I’m excited to announce that @harvey has acquired @guardrails_ai ! Back when we started Guardrails, our mission was to build infra that made the AI transformation of work more reliable and secure. Our open source framework helped establish the guardrails infrastructure category, and Snowglobe, our simulation software, helped our customers save millions of dollars in operating costs. Today, legal AI is at the frontier of the knowledge work transformation, and I’m excited to carry our mission forward by bringing everything we’ve built to Harvey. I want to thank our amazing customers, the Guardrails open source community, our team, investors and my partner-in-crime @zaydsimjee for their support throughout our journey. I’m super proud of what we built together, and can’t wait to work on the frontier of legal AI with @winstonweinberg, @gabepereyra and the incredible team they built!

There are three types of advice investors can give you: business, investor market, and personal Most investors are decent at the first two.  They’ve seen a lot of companies and they obviously know what’s attractive to the market Both skills come from pattern-matching and running in the right circles - basically spreading your time.  Both are also losing value.  AI is rewriting basically everything so patterns are less relevant, and everyone is in the same companies with the same information The third kind of advice requires the opposite - spending an extreme amount of time understanding the founders and tailoring your advice to what they’re uniquely good and bad at I’d argue this type of advice matters the most.  If founders have an outsized impact on their companies, and founders differ from each other far more than companies do, then advice tailored to the individual should have an outsized effect on outcomes This is what @krisfredrickson is best at. @gabepereyra and I first met Kris in late 2023 and he invested personally in our series B. For the past three years, he has helped us: - hire multiple key folks, because he knew who Gabe and I would work well with - walk away from what would have been a disastrous acquisition, because he knew we were looking for an easy way out of a harder problem - translate our pitch for fundraising, because he knew the company inside and out - keep going when things weren’t working, because he knew what would motivate us Most importantly, Kris was one of the first to believe in me personally, not just Harvey Very grateful to continue to partner with him

New research from @sudip_r0y and @dhruvrnaik. Harness optimization moved criterion pass rate from 67.10% to 85.92% on @harvey Legal Agent Benchmark. Post-training pushed it to 88.03%. Owning your intelligence means improving the model and the harness around it.

We’ve raised $550M at a $15.5B valuation, led by @DiffusionVC and @LightSpeedVP. We’ve crossed $400M in ARR and serve 3,000 customers, including 80% of the top 100 law firms, 20% of the Fortune 500, and half of the Fortune 10. We are investing the capital from this round into our two most important resources: people and compute. Our product is expanding into new verticals and more specialized solutions for areas like contracting, litigation, deals, and compliance. As usage grows, we’re investing heavily in inference and model training to deliver frontier legal intelligence at the best possible price. Building on Tenet, our first model post-trained on open-weight models, this round will let us scale the compute and data behind future generations of our models. We believe the most successful application-layer companies will become full-stack AI companies, building across applications, agents, and models. Becoming a full-stack AI company will help our customers own more of their intelligence and build their vision of a frontier legal organization. We’re also grateful to @Sequoia, @KleinerPerkins , @A16Z , @CoatueMgmt, @Conviction, @EladGil, Evantic, GIC, @GoldmanSachs, Sapphire Ventures and Whale Rock for participating in the round and backing that vision.