Update default models to be benchmarked continuously

[guangy10]

Summary

• Promoted Qwen3-0.6B to be the default as it's not only small enough to run quickly but also covers most of advanced changes in both etLLM and optimum-executorch.
• Removed tinyllama as nobody would care about its perf. We shouldn't use device farm for correctness testing
• Added google/gemma-3-1b-it to apple perf (private)

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/11610

• 📄 Preview Python docs built from this PR

Note: Links to docs will display an error until the docs builds have been completed.

⏳ 7 Pending, 2 Unrelated Failures

As of commit 170f15f with merge base b59f5cc ():

FLAKY - The following job failed but was likely due to flakiness present on trunk:

• pull / unittest-editable / linux / linux-job (gh) (detected as infra flaky with no log or failing log classifier)

BROKEN TRUNK - The following job failed but was present on the merge base:

👉 Rebase onto the `viable/strict` branch to avoid these failures

• pull / unittest / linux / linux-job (gh) (trunk failure)
  /pytorch/executorch/backends/vulkan/runtime/api/containers/Tensor.cpp:651:17: error: no matching constructor for initialization of 'vkcompute::api::vTensor::TextureLimits'

This comment was automatically generated by Dr. CI and updates every 15 minutes.

[guangy10]

Have to delist the google/gemma-3-1b-it from continuous benchmarking.

@kimishpatel @cbilgin FYI, Still can't run google/gemma-3-1b-it on-device due to this bug #11618. The issue wasn't caught on Android side because of another bug in the Android benchmark app here #11620.

[guangy10]

Can't merge, try rebase with no code changes