Proximal B Verified
RL environment company · Coding
- Domains
- Coding Long-Horizon Private Codebases
- On X
- @proximalhq ↗
- Funding
- $15M raised · rounds and investors
- Tier
- B why
- Snapshot
What Proximal does
Proximal is a research lab that builds complex post-training data and long-horizon coding environments for frontier models, and publishes coding and cyber benchmarks such as FrontierSWE.
- Sells
- RL environments
- Leads with
- Coding
- Based in
- San Francisco, USA
- Team
- 26-50
- Website
- proximal.ai ↗
What Proximal sells
Read from Proximal's own website, checked October 12, 2026: the line that puts it in each category.
RL environments
“How we run agentic RL across Modal GPUs and our own Kubernetes sandbox fleet.”
proximal.ai ↗“34 tasks, 13 frontier models, 20-hour budgets. Our hardest coding benchmark yet.”
proximal.ai ↗
Domains Proximal covers
Coding leads with
“Our ultra-long-horizon coding benchmark. Frontier models clear only a fraction of its tasks.”
proximal.ai ↗“Our investigation into the CyberGym benchmark and reliable cyber evaluations.”
proximal.ai ↗
Benchmarks Proximal built or runs
FrontierSWE built it
“FrontierSWE is developed by the Proximal team in collaboration with external collaborators from academia and industry.”
proximal.ai/blog/frontierswe/ ↗
Proximal competitors and alternatives
The companies that sell the most of the same things in the same domains, and what they share with Proximal.
- Datacurve Also sells RL environments and evals · also in coding and security
- Idler Also sells RL environments and evals · also in coding and security
- Revelo Also sells RL environments and evals · also in coding and security
- Abundant Also sells RL environments and evals · also in coding
- Bespoke Labs Also sells RL environments and evals · also in coding
Tier B
Growing: 11 to 50 people, or $3M+ raised.
How much stands behind a tier: usage, who backs it, a confirmed line from its own site or paper, and a clear read on X. High with three of the four, medium with two, low otherwise.
- team 11-50
- raised $15M
- valued $300M
- investors: General Catalyst, Scribble Ventures, SV Angel
- in curated RL directories (5 pts)
On X · well regarded, moderate buzz, from 14 posts about Proximal in the last month
Infrastructure partners like Modal, Baseten and Radixark highlight Proximal's post-training stack, and FrontierSWE v2 draws praise plus some confusion about model costs and rankings.
How we got to tier B
- Adoption: team 11-50; raised $15M; valued $300M; investors: General Catalyst, Scribble Ventures, SV Angel. Tier B.
- On X: well regarded, moderate buzz. “Infrastructure partners like Modal, Baseten and Radixark highlight Proximal's post-training stack, and FrontierSWE v2 draws praise plus some confusion about model costs and rankings.” No move.
- In curated RL directories: 5 points, a floor of C.
- Confirmed: a line from its own site or paper places it.
A clear view on X moves a tier one step up or down; most entries don't move. How tiers work.
Proximal funding
- Raised
- $15Msum of the rounds below
- Latest round
- Seed $15M2026-09 · led by General Catalyst
- Valuation
- $300M2026-09
Backed byGeneral CatalystChemistrySV Angel
Rounds
- Seed · $15M at a $300M valuation · 2026-09 · led by General Catalyst
We’re fortunate to be backed by General Catalyst who led our $15m at a $300m valuation
- Read: “We’re fortunate to be backed by General Catalyst who led our $15m at a $300m valuation”
- Found word for word on its own site: proximal.ai/blog/new-frontiers
- The quote states the amount and who led it
- Also named by 1 other page read
Funds 3
- General Catalyst led a round
Named in the round quoted above.
- Chemistry
as well as SV Angel, Diede van Lamoen (Go Global Ventures), Chemistry
- Read: “as well as SV Angel, Diede van Lamoen (Go Global Ventures), Chemistry”
- Found word for word on its own site: proximal.ai/blog/new-frontiers
- Also named by a curated directory
- SV Angel
as well as SV Angel, Diede van Lamoen (Go Global Ventures), Chemistry
- Read: “as well as SV Angel, Diede van Lamoen (Go Global Ventures), Chemistry”
- Found word for word on its own site: proximal.ai/blog/new-frontiers
- Also named by a curated directory
Angels 4
- Erik Bernhardsson
Erik Bernhardsson (CEO, Modal)
- Read: “Erik Bernhardsson (CEO, Modal)”
- Found word for word on its own site: proximal.ai/blog/new-frontiers
- Also named by a curated directory
- Kevin Weil
Kevin Weil (former CPO, OpenAI)
- Read: “Kevin Weil (former CPO, OpenAI)”
- Found word for word on its own site: proximal.ai/blog/new-frontiers
- Also named by a curated directory
- Liam Fedus
individuals like Liam Fedus (CEO, Periodic Labs)
- Read: “individuals like Liam Fedus (CEO, Periodic Labs)”
- Found word for word on its own site: proximal.ai/blog/new-frontiers
- Also named by a curated directory
- Diede van Lamoen
as well as SV Angel, Diede van Lamoen (Go Global Ventures), Chemistry
- Read: “as well as SV Angel, Diede van Lamoen (Go Global Ventures), Chemistry”
- Found word for word on its own site: proximal.ai/blog/new-frontiers
- No other page or directory names it yet
Every figure is quoted from its own site, its accelerator or a funding announcement; a round's amount and lead appear only when its quote states them. Startup databases are never used as evidence. Checked 2026-10-11. How funding is checked.
Training methods
Methods Proximal uses, offers or builds for, with public evidence: its own website says so, or two independent sites do.
- GRPO Group Relative Policy Optimization On its own website
“So far, we have found that using agentic judges to check the qualitative differences between rollouts leading to positive rewards and rollouts leading to negative rewards is useful: GRPO maximizes the likelihood of positive rollouts while minimizing the likelihood of negative rollouts, so it is intuitive that tasks in which positive rewards are caused by fundamentally different reasoning strategies rath…”
proximal.ai“To validate our infrastructure, we post-trained Qwen3.8-27B on an internal set of agentic coding tasks. We trained a rank-32 LoRA with 128 environments per step and a group size of 8 using vanilla Group Relative Policy Optimization ( GRPO). We categorized our tasks based on difficulty and ordered them from easy to hard, with the ability to retry tasks with no passing attempts later in training.”
proximal.ai
Where Proximal sits
Its domain on the map, live. Hover a logo for its name; click to open it. Open it on the full map
Latest from Proximal on X
@proximalhq ↗Next: browse all environments. Entry from the RL Research daily scrape of public sources, 2026-11-05 snapshot.