CoreCraft B
AI benchmark · Multi-Domain
- Built by
- Surge
- Domains
- Multi-Domain
- On X
- @HelloSurgeAI ↗
- Tier
- B why
- Snapshot
What is CoreCraft?
CoreCraft measures whether AI agents can perform multi-step, domain-specific customer support work in a simulated enterprise environment with 23 tools, scored by expert-authored rubrics.
Enterprise work mostly
“CoreCraft is a fully operational enterprise simulation of a customer support organization, comprising over 2,500 entities across 14 entity types with 23 unique tools”
arxiv.org/abs/2602.16179 ↗
Who built CoreCraft
Surge built it
“We introduce CoreCraft, the first environment in EnterpriseBench, Surge AI's suite of agentic RL environments.”
arxiv.org/abs/2602.16179 ↗
Paper: arxiv.org/abs/2602.16179 ↗
Tier B
Established: 15 or more models scored, or curated RL directories rate it.
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.
- built by Surge
- in curated RL directories (5 pts)
How we got to tier B
- Adoption: built by Surge. Tier C.
- On X: unclear, none buzz No move.
- In curated RL directories: 5 points, a floor of B. Raised to B.
- 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.
Training methods
Methods CoreCraft uses, offers or builds for, with public evidence: its own website says so, or two independent sites do.
- RLHF Reinforcement Learning from Human Feedback On its own website
“Overview: Anthropic is One of the World’s Leading AI and LLM Companies The Problem: Building Trustworthy, High-Quality Human Feedback The Solution: Rich Human Feedback via Surge AI’s RLHF Platform Results: Anthropic’s Claude – a Safe, Highly Capable, State-of-the-Art AI Assistant”
surgehq.ai“Children (and adults and AI) learn from preferences and rewards: praise a ballet dancer when her movements flow, and she’ll become better and better. The RLHF data we generate embodies the richnesses and subtleties of the world — the notes that separate the ordinary from the inspired, Salieri from Mozart.”
surgehq.ai - RLVR Reinforcement Learning with Verifiable Rewards On its own website
“To check whether ComplexConstraints measures something worth optimizing for, we built a companion training set of 1,000 examples, and trained Qwen3-4B using RLVR with a judge model.”
surgehq.ai“ComplexConstraints and Beyond: Expert Rubrics for RLVR”
surgehq.ai - GRPO Group Relative Policy Optimization On its own website
“…fic work that real jobs demand. Frontier models such as GPT-5.2 and Claude Opus 4.6 solve fewer than 30% of tasks when all expert-authored rubric criteria must be satisfied. Using this environment, we train GLM 4.6 with Group Relative Policy Optimization (GRPO) and adaptive clipping. After a single epoch of training, the model improves from 25.37% to 36.76% task pass rate on held-out evaluation tasks. More importantly, these gains transfer to out-of-distribution benchma…”
surgehq.ai“- CoreCraft (enterprise RL Environment): GRPO with rubric-based rewards moved GLM 4.6 by +11.4pp on held-out tasks and transferred out-of-distribution: +4.5pp on BFCL Parallel, +7.4pp on τ²-Bench Retail, +6.8pp on Toolathlon — while Pass³, a reliability metric requ…”
surgehq.ai
Where CoreCraft sits
Its domain on the map, live. Hover a logo for its name; click to open it. Open it on the full map
Next: browse all benchmarks. Entry from the RL Research daily scrape of public sources, 2026-11-05 snapshot.