feat(sambanova): add new models [bot]

[hganwani-droid]

Auto-generated by model-addition-agent for provider sambanova.

---

Note

Low Risk
Low risk config-only updates, but changes to context_window/max_output_tokens may affect request validation and truncate outputs for some SambaNova models.

Overview
Updates SambaNova model definition YAMLs to align published limits by adjusting context_window values (mostly to 131072, plus 8192/32768/65536 where applicable) and adding explicit max_output_tokens caps across multiple models.

Also corrects the Qwen3-235B model identifier and updates a few models’ output limits (notably reducing Meta-Llama-3.1-8B-Instruct max_output_tokens to 4096 and setting gpt-oss-120b output cap to 131072).

^{Written by Cursor Bugbot for commit 6a064e2. This will update automatically on new commits. Configure here.}

[cursor]

Cursor Bugbot has reviewed your changes and found 3 potential issues.

^{Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}