modelparams@0.0.68

Models added

• nvidia/llama-3.2-90b-vision-instruct

Parameters changed

• fireworks/inkling: updated top_p
• google/gemini-3.7-flash: updated generationConfig.thinkingConfig.thinkingLevel
• groq/gpt-oss-20b: updated reasoning_format

What's Changed

• chore(catalog): sync from gateway — 1 added, 3 corrected, 0 retired by @guillaumegay13 in #395
• chore: release modelparams@0.0.68 + modelparams-py@0.0.49 by @guillaumegay13 in #396

Full Changelog: https://github.com/mnfst/modelparams.dev/compare/modelparams-py@0.0.48...modelparams@0.0.68