This morning, we flipped the switch:
Ethan: @Zero What model are you currently powered by?
Zero: I'm powered by Claude Opus 4.7 (claude-opus-4-7).
Claude Opus 4.7, Anthropic's newest frontier model, released yesterday, is now live across vm0. Every agent in the platform can be upgraded to it today.
Here is what changed, the pricing, and how to enable it on your agents.
What Opus 4.7 brings to vm0 agents
Three capabilities in Opus 4.7 map directly to how vm0 agents work in production:
Sustained performance on long tasks. Opus 4.7 holds coherence across extended, multi-step runs — the kind of workflow vm0 agents handle daily: scheduled digests, multi-source research, document drafting across many sections. Early user reports back this up. One developer described a 9-hour overnight session with Opus 4.7 that produced 15 commits across 5 repositories and fixed a bug that had silently run for weeks.
Self-verification. The model now checks its own outputs for internal contradictions and reasoning gaps before returning them. For agents that generate customer replies, financial summaries, or status updates, this reduces the manual review burden meaningfully.
More precise instruction following. Agents with detailed system prompts — strict tone rules, output templates, conditional routing — will have those instructions respected more reliably. Less drift under complex inputs.
On independent benchmarks, document reasoning reportedly improved from 57% to 80%, and on structured coding evaluation Opus 4.7 scored 95/98 — a signal worth weighing for anyone running agents that do meaningful work in code.
Pricing: no change
Opus 4.7 is priced identically to Opus 4.6 on vm0. Same per-token rate, same credit consumption. Upgrading your workspace to 4.7 does not affect your bill in any way, you get the capability improvements at the price you were already paying.
There is no reason to hold back on upgrading for cost reasons. If Opus 4.6 fit your budget, Opus 4.7 does too.
How to enable Opus 4.7 for your workspace
The model switch is made at the workspace level, one change applies to every agent using the default built-in model.
Step 1. Open the workspace manager.
In the top-left of your vm0 workspace, click Manage next to your organization name.
Step 2. Go to Model Providers.
In the dialog that opens, select Model Providers from the left sidebar (under the Configuration section). On the right, find the Built-in model card and click Edit.
Step 3. Select Claude Opus 4.7.
In the Edit workspace Built-in model dialog, open the Select model dropdown and choose claude-opus-4-7. Click Save changes.
That's it. Every agent running on the built-in model will use Opus 4.7 on its next invocation.
Verify the change. You can confirm any agent's active model directly in Slack:
@[AgentName] what model are you currently powered by?
When to upgrade, and when to wait
Opus 4.7 is the right choice for agents where output quality directly shapes decisions — research, financial review, customer communication, product writing, code review.
For lighter, high-volume tasks where speed and cost matter more than reasoning depth, Sonnet remains a strong default. If you are not sure which model fits a given agent, ask Zero for a recommendation based on the agent's current configuration and workload.
Opus 4.7 is available now on all vm0 plans. No additional setup required, just select the model and your agents are ready.
Questions or feedback? Reach us at support@vm0.ai.
