Agentic harness comparison
Jules vs OpenAI Codex
Compared on where each one runs, which editors it supports, how autonomous it is, and how it bills — the four things that actually decide Jules vs OpenAI Codex. Rankings reflect fit and trade-offs, not a single universal winner; partner links, when present, do not replace that methodology.
Details last verified September 3, 2026
Jules
Cloud async agent (web + GitHub) · Agentic
OpenAI
OpenAI Codex
Terminal and cloud · Agentic
| Specification | Jules | OpenAI Codex |
|---|---|---|
| Vendor | OpenAI | |
| Runs in | Cloud async agent (web + GitHub) | Terminal and cloud |
| Editor support | Browser workspace, GitHub (opens PRs) | Any (editor-agnostic) |
| Autonomy | Agentic | Agentic |
| Pricing model | Hybrid | Hybrid |
| Requires changing editor | No | No |
| SWE-bench Verified (Gemini 2.5 Pro, async agent) (2026-03-01) | 51.8% | Not verifiedUnverified |
Benchmark scores carry the month they were measured. Values we could not confirm are shown as “Not verified” rather than estimated.
Benchmark receipts
Jules · SWE-bench Verified (Gemini 2.5 Pro, async agent): source
Jules
Google's asynchronous coding agent: it clones your repo into a secure cloud VM, writes a plan you approve, executes multi-file changes in the background, and opens a reviewable pull request — a direct competitor to Codex and Devin Cloud.
Best for
- Delegated async coding
- Open-source backlog triage
- Teams on Google AI subscriptions
Watch out
Bundled with Google AI tiers (verified 2026-08-29, jules.google and blog.google): Free 15 tasks/day, 3 concurrent; Google AI Pro $19.99/mo (~100 tasks/day, 15 concurrent); Google AI Ultra $199.99/mo (300 tasks/day, 60 concurrent — the $124.99 figure was the first-3-months 50% intro promo, not the list price). No public API. Launched Aug 2025. SWE-bench Verified ~51.8% (Gemini 2.5 Pro) is from one secondary source — treat as indicative, not a Google-published figure.
OpenAI Codex
OpenAI's delegated coding agent available in the CLI, IDE extension, and cloud, for teams already standardized on OpenAI billing and tooling.
Best for
- Delegated coding tasks
- Existing OpenAI commitments
Watch out
ChatGPT plans include Codex with plan-based limits; API-key login is pure usage billing and does not unlock cloud features (cloud review/Slack). Pricing (verified 2026-08-29, chatgpt.com/codex/pricing and developers.openai.com/codex/pricing): Free, Go, Plus $20/mo, Pro 5x $100/mo, Pro 20x $200/mo, Business and Enterprise. On April 2, 2026 Codex moved from per-message to token-based credit billing; usage is metered by model at API token rates. Passed 20M+ users in Aug 2026 (OpenAI gave every plan a free one-off 'banked reset' of allowances to mark it), and the GPT-6-Astra model catalog arrived via backport in CLI 0.153.1 (2026-09-03, API-configurable rather than in the picker). One cloud task can consume many credits, so size by the 5-hour rolling window plus weekly caps, not a flat seat count.
Common questions
Jules vs OpenAI Codex
Answered from the verified figures on this page rather than general guidance.
Do I have to change editor to use Jules or OpenAI Codex?
Neither does. Jules runs in Browser workspace, GitHub (opens PRs) and OpenAI Codex runs in the terminal and cloud alongside any editor, so people keep the setup they already have.
How are Jules and OpenAI Codex priced?
Both are hybrid. The headline seat or plan price is not the heavy-agent bill — included credits, shared pools, and overages dominate once you loop agents hard. Estimate cost per successful task and set spend caps.
Is Jules or OpenAI Codex better for large refactors?
Both are agentic, so neither has a structural advantage on delegated work. Judge them on editor fit and price instead.
Should I run Jules and OpenAI Codex together?
Often yes as a hybrid stack rather than forcing one winner. Keep the stronger in-editor assistant for daily tab/chat work and the more autonomous surface for longer delegated tasks, then standardise on a portable repo rules file so project memory survives switching. Pilot both on the same practice repository before consolidating seats.
Will my project rules transfer between Jules and OpenAI Codex?
Usually not automatically. Cursor rules, CLAUDE.md, AGENTS.md, and extension-specific configs are different conventions. Keep one source of truth in the repository (commonly AGENTS.md) and copy or link into each tool’s native format so switching does not wipe months of project teaching.
Is Codex still billed per seat or per message?
No. ChatGPT plans include Codex with plan limits/credits, and API-key login bills Platform tokens separately. A single agentic task can burn many credits — do not budget as if each message were one flat charge.
Do coding agents replace code review?
No. Agents can draft patches faster, but they do not remove the need for human review of plans, tool calls, permissions, diffs, tests, and rollback paths. Treat autonomy as a reason for stronger review, not less — see the coding assistant rollout guide at /ai-harnesses/guides/coding-assistant-rollout-guide.
Should we harden MCP before rolling out a coding agent?
Yes for any agent that can install MCP servers or run tools. Maintain a short allowlist, review each server’s folder and network access before install, fail closed on unknown tools, and log tool calls. .cursorignore and similar rules do not block MCP or shell access once a command is approved.
Next step
Which one fits
Autonomy and editor constraints decide this far more reliably than a benchmark score.
If most of your week is small, frequent edits, an assistive workflow beats delegation and an agentic harness sits idle while billing. If you regularly hand over work spanning many files, the reverse is true.
Where one tool requires changing editor and the other does not, that difference usually outweighs any capability gap across a team — rollouts fail on adoption, not capability.
“Open vendor” links use a first-party redirect for click counting. They are not commissionable partner placements unless a programme is active for that vendor. Affiliate disclosure.