Add gpt5 benchmark tool preset support

[enyst]

Summary

• add gpt5 to the benchmark tool preset enum and common CLI choices
• route benchmark tool-preset dispatch to the existing GPT-5 preset implementation
• add focused coverage for parser acceptance and preset dispatch helpers

Why

The GPT-5 preset already exists in software-agent-sdk here:

• https://github.com/OpenHands/software-agent-sdk/blob/main/openhands-tools/openhands/tools/preset/gpt5.py

This PR adds the benchmark-side support needed to accept --tool-preset gpt5 so we can run evals with that preset.

Testing

• uv run pytest tests/test_tool_presets.py -q
• uv run pre-commit run --files benchmarks/utils/args_parser.py benchmarks/utils/models.py benchmarks/swebench/run_infer.py benchmarks/swebenchmultilingual/run_infer.py benchmarks/hybridgym_funclocalize/run_infer.py benchmarks/hybridgym_depsearch/run_infer.py benchmarks/hybridgym_funcgen/run_infer.py benchmarks/hybridgym_issuelocalize/run_infer.py tests/test_tool_presets.py

---

This PR was created by an AI assistant (OpenHands) on behalf of the user.

[all-hands-bot]

Clean, well-structured additive change that follows existing patterns. Tests are appropriate and focused.