Default BrowserCode evaluations to 100 concurrent tasks

[MagMueller]

BrowserCode V2 runs now default to up to 100 concurrent tasks in both run_eval.py and GitHub Actions. The workflow accepts concurrency values from 1 to 200, and the README documents overrides and GitHub account limits. The benchmark still selects all 200 tasks; legacy V1/Stealth concurrency is unchanged.

Validation: 66 unittest cases completed successfully (two optional binary checks skipped), Ruff, actionlint, and diff checks passed. Verified CLI overrides and workflow concurrency boundaries. This changes configuration defaults; no new 100-concurrent live benchmark was dispatched.

---

Summary by cubic

Defaults BrowserCode V2 evaluations to up to 100 concurrent tasks in both run_eval.py and the GitHub Actions workflow, while keeping the benchmark set at all 200 tasks. Legacy V1 and stealth concurrency defaults are unchanged.

• The bcode executor now defaults to 100 concurrent tasks; other executors keep the prior default of 3.
• The workflow's max_parallel input now defaults to 100 (was 12) and accepts values from 1 to 200 (was 1 to 50).
• README documents the new default and how to override it with --parallel or max_parallel.

^{Written for commit 66aec34. Summary will update on new commits.}

[cubic-dev-ai]

All reported issues were addressed across 4 files

<sub>Heads up: you’re close to your flex budget. Increase your flex budget so reviews don’t pause.

Fix all with cubic | Re-trigger cubic</sub>