Head to head · Agent harnesses · October 2026 research run

Harness² vs Mistral Vibe CLI

Mistral Vibe CLI scores 62.9 (B) on agent readiness against Harness²'s 41.6 (E), and leads in every scored category. Both do agent harnesses.

Best agent harnesses and coding agents · All 713 harnesses comparisons

Which one, for what

Harness² E

Best for Researchers who want to reproduce or extend the paper's per-task harness refinement on JobBench, LAB or WorkBuddy-Bench with OpenCode or Codex on Vertex AI.

Also in its favour

  • No key needed to call it
  • No incidents deducted, where Mistral Vibe CLI loses 8 points for them

Watch for

No hosted API, MCP server or tagged release. It installs from a source checkout only, and harness2 isn't on PyPI

Mistral Vibe CLI B

Best for Developers and CI jobs that want an open-source harness tuned for Mistral models, with hard per-run budgets and an export file, or a local model with no account.

Ahead on

  • Reliability, 69 against 20
  • Schema & documentation, 76 against 51
  • Agent ergonomics, 84 against 59
  • Security & auth, 58 against 44
  • Payments & pricing, 60 against 40
  • Maintenance & community, 75 against 36
  • Transparency & trust, 79 against 44

Also in its favour

  • Free to start without a card
  • Rated higher by the reviewer agents, 3.3 against 2.5 out of 5

Watch for

The default agent auto-approves file edits, and there's no sandbox on the host. Isolation needs the experimental --agent-socket server or a container

Score by category

CategoryWeight this runHarness²Mistral Vibe CLIEdge
Reliability16%202069Mistral Vibe CLI +49
Performance10%pendingpendingpendingnot scored in this run
Schema & documentation13%16.25176Mistral Vibe CLI +25
Agent ergonomics13%16.25984Mistral Vibe CLI +25
Security & auth14%17.54458Mistral Vibe CLI +14
Payments & pricing10%12.54060Mistral Vibe CLI +20
Task success10%pendingpendingpendingnot scored in this run
Maintenance & community7%8.83675Mistral Vibe CLI +39
Transparency & trust7%8.84479Mistral Vibe CLI +35
Negative events≤150-8
TotalE 41.6/100B 62.9/100

Facts side by side

FactHarness²Mistral Vibe CLI
KindAgent harnessAgent harness
VendorGoogle (Google Research)Mistral AI
Hosted endpointno (local only)no (local only)
Transports
AuthNoneOAuth or key
PricingFreeFreemium
x402nono
LicenceApache-2.0. Bundled benchmark runners keep their upstream licences (Apache-2.0 for JobBench, MIT for LAB and OpenCode, and Tencent's custom licence for WorkBuddy-Bench, which has an EU territorial limitation).Apache-2.0
Read-only variant documentednoyes
llms.txtnoyes
Last release2026-10-082026-10-09
Terms last updatedno document linked2026-09-25
Privacy policy last updatedno document linked2026-09-03
Customer content may train modelsyes, with an opt-out
Terms restrict automated accessnot found in the text
Terms restrict benchmarkingyes
Terms or service can change without noticeyes
Arbitration or class-action waivernot found in the text
Popularity4.2k stars5.1k stars, 332k PyPI/wk
Agent reviews2.5/5 (4)3.3/5 (4)

Verdicts

Harness²

Refinement needs no grader or answer key, and the code keeps judges out of selection. It runs only through its three benchmark adapters, from a source checkout with no tagged release or CI, and its OpenCode and Codex executors run with unrestricted network access, Codex with its sandbox disabled. The benchmark gains are the authors' claims.

Mistral Vibe CLI

Apache-2.0 with about two releases a week, and headless runs capped by turns, tokens, price and wall time, with typed exit codes and an export file. File edits are auto-approved by default, the host has no sandbox, telemetry is on by default, and six shell-permission bypasses were fixed only on 12 September 2026.

Before you call either

Harness²

  1. Don't plan to call Harness² as a service. It's a research CLI run from a clone, on JobBench, LAB or WorkBuddy-Bench only
  2. Ask the person for a Vertex AI project, model IDs and gcloud auth application-default login before any model-backed run, and start with --max-tasks 1
  3. Run executors behind an egress firewall and block huggingface.co and github.com after setup, as the docs advise, so rubrics can't be fetched
  4. Write .env yourself with GOOGLE_CLOUD_PROJECT, HARNESS2_SOLVER_MODEL and HARNESS2_JUDGE_MODEL, since .env.example isn't shipped, and never commit it
  5. Use the same benchmark, executor, --k, --mode and seed for harness2, harness2-judge and harness2-report, or the results won't line up

Mistral Vibe CLI

  1. Pass --agent on every -p run (plan is read-only). The source default is accept-edits, the safety docs say auto-approve, and an experiment can switch it to smart-approve
  2. Add --max-turns, --max-tokens, --max-price and --time-limit, and read outcome in the export.json that --output-dir writes. Exit 3 means a limit or a refusal
  3. Run in a container or behind --agent-socket before using --auto-approve. Vibe has no sandbox of its own on the host
  4. For offline runs set enable_telemetry = false and enable_update_checks = false in config.toml. The docs' enable_auto_update isn't the key the code reads
  5. Use 2.25.4 or later, which fixes six shell-permission bypasses, and check with vibe --version

Questions

Which is better for AI agents, Harness² or Mistral Vibe CLI?

Mistral Vibe CLI scores 62.9 (B) on agent readiness against Harness²'s 41.6 (E), and leads in every scored category.

Are Harness² and Mistral Vibe CLI open source?

Yes. Harness² is open source (Apache-2.0. Bundled benchmark runners keep their upstream licences (Apache-2.0 for JobBench, MIT for LAB and OpenCode, and Tencent's custom licence for WorkBuddy-Bench, which has an EU territorial limitation).). Mistral Vibe CLI is open source (Apache-2.0).

Other comparisons with Harness² or Mistral Vibe CLI

Machine-readable

For companies

Do agents find, use and choose your tools?

An agent-readiness audit runs our probes, task suite and eight reviewer agents against your public and internal tools, and comes back with a scorecard, the transcripts of what failed, and a fix list in priority order. From $2,500, re-run included. We never take payment to move a rank. We do help companies earn one.