Runner: update the seed_options dictionary

[finozzifa]

The option "KN_PARAM_MS_SEED": 1066, caused linopy Knitro interface to fail.

The changes proposed by @KristijanFaust-OET in the PR #433 solve it.

Checklist

For PRs adding new benchmark instances:

• I consent to releasing these benchmark instance files under the CC BY 4.0 license
• The benchmark name and size instance name follow the conventions indicated in the template
• I have tested that this model instance can be solved to optimality in [time] with [solver] solver, using [options], on a [spec] machine

For Benchmark team:

• Upload the LP/MPS files (compressed using gzip -9 <filename>) to our GCS bucket
• Run benchmarks/categorize_benchmarks.py on them to obtain problem stats and size category
• Run tests/validate_urls.py to ensure URLs are consistent with benchmark and size instance name
• Update the summary table in the Key Insights page (if the proposed benchmark(s) is/are adding any new features)
• Test that some solver solves these benchmarks within our timeouts on our infra
• Add CC BY 4.0 license to the metadata file

For changes to the website:

• I have tested my changes by running the website locally

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
solver-benchmark	Ready	Preview, Comment	Feb 25, 2026 2:52pm

[siddharth-krishna]

As discussed, let's make sure that the main branch only sets the random seed and the mip gap (or equivalent) option for each solver. (Btw we're missing mip gap for knitro, could you please add?) Thanks!

[finozzifa]

Closed as not necessary anymore!