Harvey

AI for the world’s most complex legal work.

[Harvey @harvey]

Would Harvey Specter use Harvey?

@GabrielMacht had to ask.

[Gabe Pereyra @gabepereyra]

Q2 recap for @harvey
- +$100M NNARR
- 53% DAU/MAU

Key hires (including Q1)
- Anique (CPO) - prev VP of Product at Rippling
- Rachel (CMO) - prev CMO at Notion
- Brooks (CISO) - prev CISO at Roblox
- Keith (CSO) - prev CPO at Google

Product
- Agent unification - cloud agents can use all Harvey product surfaces
- Command center (EA) - monitor adoption and ROI by use case
- Contract intelligence (EA) - agentic contracting platform for enterprises

Eng
- Migration to cloud agent infrastructure
- Integrating open source inference providers
- Scaling document processing (54TB / week)

AI
- Legal Agent Bench
- Open source post training
- Published multiple research directions with partners

We invested heavily in cloud agent infrastructure at the end of last year and in Q1. In Q2 we also unified many of our product surfaces (collapsed as @winstonweinberg says) by making them all tools accessible by our cloud agents.

Prior to this, there were a lot of capabilities in Harvey that were often only discovered by power users. As cloud agents get better and our product becomes more connected we are seeing users discover more of the product by learning from their agents (see plot of product surfaces per user).

[Niko @nikogrupen]

One of the underrated pieces of our post-training work with @baseten is exploring KV cache compaction.

In our training runs over Legal Agent Bench, agents converged to similar policies, roughly:
1. Ingest client matter documents up front
2. Analyze + iterate on intermediate work product
3. Draft final work product
4. Revise

For heavier matters (real diligence datarooms can easily be 10M+ tokens), step (1) becomes the bottleneck.

Natural language compaction is the logical starting point for long context problems, and we’ve shown a few times that NL compaction improves performance, but at 10M+ tokens even NL summaries get unwieldy.

NL compaction is also lossy – the model is forced to compress a rich internal state into prose. Over long-horizon trajectories, subtle degradations of compacted context can compound. In legal, if you lose a key clause, date, or defined term the impact on work product is catastrophic.

KV cache compaction is an interesting alternative. STILL, a method from Baseten Research, compresses the KV cache directly. For each layer, learned latent queries cross-attend over the full cache and write a much smaller set of synthetic keys and values in a single forward pass. The compressed state stays in a representation the model can attend to directly, rather than being forced through a text bottleneck.

This has two cool properties
1. Because compression lives in the model's latent space, superposition lets each slot store significantly more information than a discrete token can. That gives the agent much denser working context than a memo or summary.
2. The compactor can be trained in-domain, so it can learn what kinds of details legal tasks actually need to preserve.

Having the right memory primitives is increasingly part of the post-training equation for legal work, and knowledge work more generally. Much more to come here and s/o @oneill_c, @mudithj, and team for the collaboration and innovative work here.

[Harvey @harvey]

Learn more: https://t.co/kjlLnEWdHV

[Harvey @harvey]

Now live in Harvey: Claude Sonnet 5.

Sonnet 5 scored 5.8% all-pass on our Legal Agent Benchmark (LAB), building on Sonnet 4.6 with broad gains in legal accuracy and output quality.

[Christopher Sim @thisiscsim]

A few photos from design night with @harvey and @DecagonAI. Thank you to everyone who stopped by ✨

[Harvey @harvey]

Inference-time model routing based on legal practice area dramatically improves agent performance.

So do other types of "blended intelligence":
- collaborative model teams,
- advisor-executor patterns,
- and model routing based on inferred user preference.

While these methods appear promising, they come with substantial risks and eval challenges.

We're experimenting with all of them at Harvey. Read more from our Head of Legal Research @ItsJulioPereyra:

[Julio Pereyra @ItsJulioPereyra]

https://t.co/L9UcXYwvtA

[Harvey @harvey]

Learn more: https://t.co/b2a1MERsIY

[Harvey @harvey]

Harvey is now the exclusive Legal AI partner of the Toronto Maple Leafs.

We're proud to partner with one of the most iconic franchises in sports as we continue investing in Toronto and expanding our presence across Canada.

[Harvey @harvey]

Harvey FORUM comes to Sydney, Australia, September 2, 2026.

One day in Sydney, bringing together leaders from across APAC for expert perspectives, peer dialogue, and the conversations reshaping legal.

Save the date: more details coming soon.

[Harvey @harvey]

@appliedcompute Check out @appliedcompute's agent-training deep dive thread: https://t.co/OJ7wOnVdJt

[Applied Compute @appliedcompute]

We partnered with @harvey to post-train the state-of-the-art legal agent on their LAB benchmark. It surpasses Opus 4.8 Max and GPT-5.5 xhigh.

[Harvey @harvey]

We partnered with @appliedcompute to train a legal agent.

We optimized each part of the agent stack:
- the evaluation loop
- the agent harness
- and post-trained the underlying GLM-5.1 model.

The result? The agent achieved the highest rubric pass rate on our Legal Agent Benchmark (LAB) of any available model.

Much more in our agent-training deep dive with @appliedcompute:

[Harvey @harvey]

https://t.co/hlb6DJ8zbV

[Harvey @harvey]

https://t.co/hlb6DJ8zbV

[Molly O’Shea @MollySOShea]

NEW: Harvey Co-Founder + Head of Applied Research on the *Token Reckoning*

Valued at $11B, @harvey is on a mission to win the entire legal category, competing head-on against the trillion-dollar labs

Coding agents hit Karpathy's "agents work now" inflection in late 2025. Harvey Co-Founder @gabepereyra (fmr Google Brain, DeepMind & Meta) argues legal is hitting its version of that curve right now.

With both Gabe + Head of Applied Research @nikogrupen, we cover:
- Open-sourcing LAB (legal agent benchmark): 1,200+ tasks across 24 practice areas, 75,000+ rubric criteria
- Who's leading the leaderboard
- Harvey is the largest embeddings consumer for some of the labs
- Why every law firm has to be multi-model: conflict risk
- The billable hour is coming back, this time for AI tokens

FYI: Harvey Labs is the internal research group pushing the frontier of legal AI. Run by Niko (fmr multi-agent RL at Google Brain) & @ItsJulioPereyra (fmr clerk + Big Law attorney), it partners with the labs, research community, & academia to bring frontier agent research into Harvey.

𝐓𝐈𝐌𝐄𝐒𝐓𝐀𝐌𝐏𝐒
(00:00) Gabe Pereyra (Co-Founder) & Niko Grupen (Head of Applied Research)
(00:50) Inside Harvey's legal agent Benchmark
(05:10) What happens after Benchmarking?
(06:37) Why Harvey open sourced its research
(09:21) Training models without client data
(10:32) Google Brain vs. DeepMind
(12:34) From Researcher to Founder
(15:15) The Rise of the Inference Layer
(18:38) The Agentic Shift
(21:16) Harvey's 13 trillion tokens
(23:48) AI's Biggest cost misconception
(28:37) How Top AI founders learn
(31:52) Learnings from Jensen Huang
(34:14) How Harvey finds talent
(35:41) Niko on Harvey's breakthroughs
(36:38) Building a legal dataset from scratch
(38:32) How to read AI Benchmarks
(39:51) Niko's research playbook
(40:51) The Opportunity beyond Benchmarks
(41:45) Why Agent Harnesses matter
(43:04) The Rise of Organizational AI

[Harvey @harvey]

Learn more: https://t.co/IFvma74WId

[Harvey @harvey]

Visualize legal data on demand in Harvey.

Ask Harvey to build interactive timelines, entity charts, and compliance dashboards from your inputs - filterable, shareable, and ready in minutes.

[Harvey @harvey]

We're training the first in a series of legal foundation models.

Harvey Labs is hiring across the post-training, agent, and data stack. We're also interested in acquiring teams and neolabs with talent in these areas.

DM @gabepereyra if you're excited by frontier legal AI.

[Gabe Pereyra @gabepereyra]

Model strategy for @harvey:

We are working on the first model in our legal foundation model series, inspired by @cursor_ai's Composer. Two goals:

1. Allow us to serve frontier intelligence across our product surface areas at an affordable price and a strong security posture.
2. Create the foundations for law firms to build their own specialized models and own their own intelligence.

The model series will focus on complex client matters that span months and take dozens of associates. The agentic system will learn to control legal tech tools, sub agents and ask for help from frontier models or human partners, much like a senior associate.

We’ve open sourced benchmarks for evaluating our initial post training work that represents work done by associates and in-house lawyers. We are scaling these significantly using synthetic and human pipelines as well as building private evals for firms.

Open sourcing this data has allowed us to quickly validate the feasibility of post training open weight models for legal work. With our research partners we’ve already shown promising results post training open source models to approach frontier performance:

1. @baseten - novel compaction strategies for analyzing large data rooms.
2. @FireworksAI_HQ - matching frontier performance by using frontier as an advisor.
3. @appliedcompute - improving performance and reducing cost of large scale review tables.
4. @trajectorylabs & @nvidia - sovereign continual learning over client matters.

We plan to continue to invest heavily in working with research partners and open sourcing our data, models and research as much as possible. We believe open research in legal will be important to building trust in the frontier ecosystem.

We are also scaling our research team. Harvey Labs is our internal research group, responsible for pushing the frontier of legal intelligence and working closely with labs, research partners, and academia to bring the frontier of agent research into Harvey.

Labs is run by @nikogrupen and @ItsJulioPereyra - Niko worked on multi-agent RL at Google Brain and Julio clerked and worked in BigLaw. We believe this pairing is crucial for building frontier legal AI systems. Together they have already made significant progress in scaling our data and training efforts.

The long term goal of Harvey Labs is to contribute to the research and infrastructure required for the legal industry to create a frontier ecosystem. We believe that the best version of legal super intelligence is one where each law firm, enterprise and government owns their own specialized version.

We are hiring for Harvey Labs across the post training, agent and data stack and open to acquiring talented teams / neolabs in this space. If interested please DM me.

[Harvey @harvey]

Vals AI has released a live leaderboard of our Legal Agent Benchmark (LAB).

We're excited to see LAB more deeply embedded in the eval ecosystem.

See @ValsAI's live leaderboard here: https://t.co/ZSSM7SHXFL

[Vals AI @ValsAI]

We are releasing a live leaderboard for @harvey's Legal Agent Benchmark on Vals AI.

We are the first third-party to host this benchmark live. Results are on the private, held-out test set, not the public set.

[Box @Box]

The Box MCP connector is now available in @Harvey's Connector Library to early access customers.

Search files, query documents, extract metadata, and create content in Box directly from Harvey, with existing permissions and security controls enforced throughout.

Watch it in action.👇

[Harvey @harvey]

Register here: https://t.co/GiThXaQJq4

[Harvey @harvey]

Join us for Agents Happy Hour in SF on Tuesday, June 23, co-hosted with our friends at @browserbase and @reductoai.

We’re bringing together engineers and founders for drinks and demos on building useful agents.

Food and drinks provided, space is limited. RSVP below.