CoreWeave Forge
One product to run, observe, evaluate and post-train agents (W&B, OpenPipe, sandboxes).
Updated
✨ Try AgentScout's free AI image generator →Good for: evaluating agents, running code safely, training and fine-tuning agents
Use cases
About CoreWeave Forge
CoreWeave Forge is listed in our evals & observability category and is aimed at builders. One product to run, observe, evaluate and post-train agents (W&B, OpenPipe, sandboxes). Pricing, as the maker states it: Free $0/mo; Pro from $60/mo; Enterprise custom. Maturity: Newly launched. We last checked the official link on 2026-10-18.
What is CoreWeave Forge?
CoreWeave Forge is a platform for building and improving models and agents, launched on September 30, 2026. It covers the whole loop of running, observing, curating, improving and evaluating in one connected environment. It combines Weights & Biases Models, OpenPipe's post-training expertise and marimo-based notebooks with CoreWeave's own services, and it works with other models, frameworks and clouds.
At launch it included ARIA, a coding agent that analyzes experiment and agent-observability data and proposes next steps. It also included Agent Lens for observing production agents, Sandboxes for isolated runs, Registry for versioned checkpoints and agent configs, Post-Training (serverless RL and fine-tuning) and Inference. The Free plan costs $0 a month. Pro starts at $60 a month with a 30-day trial and is meant for teams with fewer than 50 employees. Enterprise is custom.
Sources: coreweave.com/news/coreweave-forge-launches-turning-the-ai-loop-produc, coreweave.com/forge-pricing and our directory listing, checked 2026-10-18.
Who is CoreWeave Forge for?
ML and agent teams that want tracing, evals, sandboxes and post-training in one place, including existing Weights & Biases users.
Key features
- Agent tracing, evaluations and scorers
- Agent Lens observability for production agents
- Sandboxes for agent tool use, RL and evals
- ARIA agent that proposes next experiments
- Serverless RL, fine-tuning and distillation
- Free $0/mo; Pro from $60/mo; Enterprise custom
CoreWeave Forge FAQ
What is CoreWeave Forge?
CoreWeave Forge is a platform for building and improving models and agents, launched on September 30, 2026. It covers the whole loop of running, observing, curating, improving and evaluating in one connected environment. It combines Weights & Biases Models, OpenPipe's post-training expertise and marimo-based notebooks with CoreWeave's own services, and it works with other models, frameworks and clouds.
Who is CoreWeave Forge for?
ML and agent teams that want tracing, evals, sandboxes and post-training in one place, including existing Weights & Biases users.
Is CoreWeave Forge free?
CoreWeave Forge's pricing, as the maker states it: Free $0/mo; Pro from $60/mo; Enterprise custom. Check the official site for current plans.
How much does CoreWeave Forge cost?
Free is $0/month. Pro starts at $60/month (billed monthly, 30-day free trial) for teams with fewer than 50 employees. Enterprise is custom-priced.
What are alternatives to CoreWeave Forge?
Similar tools include OpenLIT, Agent Memory Leaderboard, Kore.ai Autoloop, agent-qa (Vostride), Basalt and BrowserGym.
CoreWeave Forge ratings and reviews
Be the first to rate CoreWeave Forge: tap 1 to 5 stars below. A written review is optional.
CoreWeave Forge alternatives
Other evals & observability tools in our directory. See all evals & observability tools →
OpenLIT
Open-source platform to trace, evaluate and guard agents and LLM apps.
Agent Memory Leaderboard
Public leaderboard and open evaluation protocol, run by a group of universities and labs, for comparing long-term memory systems for agents in recurring submission rounds.
Kore.ai Autoloop
Optimization engine in Kore.ai Agent Platform Artemis that evaluates agents against business goals, diagnoses failures from traces, applies targeted fixes, and re-verifies; GA Oct 2026.
agent-qa (Vostride)
A self-improving AI QA agent with persistent memory that tests web, Android and iOS apps from plain-English test steps, for developers and coding agents.
Basalt
Captures customer behavior so your AI agent can learn and improve from real interactions.
BrowserGym
ServiceNow's Gym environment for building, testing and evaluating web agents across browser benchmarks.
Related lists
🏆 See also: Best AI Agent Observability & Eval Tools · All best-of lists
🧭 Task guides: AI eval and benchmark tools
Badge for makers
Built CoreWeave Forge? Add this badge to your README or site. It links back to this page. About badges
