Who contributed to Google on August 6, 2026?

10 developers shipped this update, including Kanglan Tang, copybara-service, mattjj, release-please[bot], Matthew Tang, Google Team Member, release-please, and sofisl, and 2 more.

What were the notable Google updates?

[CI] Skip profiler tests on all free-threaded Python versions, [Mosaic GPU] Define the `b6x16_p32` MLIR dtype, and [Pallas/Mosaic GPU][NFC] Deduplicate Pallas and Mosaic GPU compilation pipelines.

Google

JAX, the GenAI SDK, and the Cloud libs - Google's open source layer

Pick a date

Topics: AI / ML Python Full archive →

The Wire · Showcase

JAX PINS PYTHON 3.14T, GENAI GAINS ROBOTICS MODEL AND ERROR HANDLING

By RepoJournal · Filed 06:59 UTC on August 6, 2026 · About Google

10 people shipped this

copybara-service @copybara-service 3 cited

Heba Alazzeh @hebaalazzeh 2 cited

Kanglan Tang @kanglant 1 cited

release-please[bot] @release-please[bot] 1 cited

Google Team Member @google-genai-bot 1 cited

ohmayr @ohmayr 1 cited

mattjj @mattjj

Matthew Tang @matthew29tang

release-please @release-please

sofisl @sofisl

JAX disabled profiler tests on free-threaded Python to unblock CI, while python-genai 2.17.0 ships the Gemini Robotics ER 2 Preview and fixes tool call limits.

JAX's profiler test now skips on all free-threaded Python versions [1] after hitting timeout walls on Python 3.14t CI. The same release cycle saw heavy Mosaic GPU work land: a new b6x16_p32 MLIR dtype [2], unified Pallas and Mosaic compilation pipelines [3], and debug dumps for layout inference [4] via MOSAIC_GPU_DUMP_MLIR_PASSES. Python-genai 2.17.0 [5] adds the Gemini Robotics ER 2 Preview model and introduces TOO_MANY_TOOL_CALLS to the FinishReason enum [6], addressing a gap when models exhaust their tool budget. On the cloud-python side, a proto-plus message fix landed in api-core [7] that restores UUID auto-population by switching from identity checks to truthiness checks, unblocking downstream libraries. The import profiler script also got faster [8] by calculating codebase statistics only once instead of on every iteration across 22 test runs. Google-cloud-python bumped api-core to 2.28.0 [9] to consolidate version parsing logic and reduce duplication.

Action items

→ Update google-cloud-python to api-core 2.28.0 before next release cycle googleapis/google-cloud-python [plan]
→ Pin python-genai to 2.17.0 if you use tool calling with many models googleapis/python-genai [plan]
→ Monitor JAX free-threaded Python support for 3.14+ stability google/jax [monitor]

References

[1] [CI] Skip profiler tests on all free-threaded Python versions google/jax
[2] [Mosaic GPU] Define the `b6x16_p32` MLIR dtype. ↗ google/jax
[3] [Pallas/Mosaic GPU][NFC] Deduplicate Pallas and Mosaic GPU compilation pipelines. ↗ google/jax
[4] [Pallas/Mosaic GPU] Allow dumping module before and after layout inference. ↗ google/jax
[5] v2.17.0 ↗ googleapis/python-genai
[6] feat: Add TOO_MANY_TOOL_CALLS to FinishReason enum. googleapis/python-genai
[7] fix(api-core): use truthiness check in setup_request_id to support proto-plus messages (#18000) googleapis/google-cloud-python
[8] fix(ci): skip loaded lines calculation on iterations after the first ↗ googleapis/google-cloud-python
[9] chore: bump google-api-core to 2.28.0 ↗ googleapis/google-cloud-python

Quick answers

What shipped in Google on August 6, 2026?: JAX disabled profiler tests on free-threaded Python to unblock CI, while python-genai 2.17.0 ships the Gemini Robotics ER 2 Preview and fixes tool call limits. In total, 42 commits, 37 pull requests, and 11 releases landed.
Who contributed to Google on August 6, 2026?: 10 developers shipped this update, including Kanglan Tang, copybara-service, mattjj, release-please[bot], Matthew Tang, Google Team Member, release-please, and sofisl, and 2 more.
What were the notable Google updates?: [CI] Skip profiler tests on all free-threaded Python versions, [Mosaic GPU] Define the `b6x16_p32` MLIR dtype, and [Pallas/Mosaic GPU][NFC] Deduplicate Pallas and Mosaic GPU compilation pipelines.

JAX ADDS NUMPY 2.6 TOP_K, INTEL GPU SUPPORT LANDS IN CI

JAX shipped native top_k alignment with NumPy 2.6.0 while simultaneously landing full OneAPI CI infrastructure for Intel GPU testing across the entire stack.

ai 59 shipped

Google Aug 4

JAX HARDENS GPU KERNELS WHILE CLOUD PYTHON PATCHES SUPPLY CHAIN RISK

JAX shipped critical Mosaic GPU optimizations that squeeze performance out of transposed memory transfers, while the cloud SDK locked down six open security vulnerabilities in sqlalchemy-spanner.

ai 100 shipped

Google Aug 3

JAX SYNCS XLA DEPENDENCY TO LATEST REVISION

JAX pinned its XLA dependency to a fresh upstream revision overnight, pulling in compiler optimizations and fixes across the ML stack.

ai 1 shipped

Google Aug 2

JAX AVERTS BREAKING CHANGE IN CUSTOM_GRADIENT API

JAX pulled back from making the `fun` parameter positional-only in custom_gradient, preserving backward compatibility just before release.

ai 3 shipped

Elsewhere on the wire

AI Agents Aug 6

LANGCHAIN HARDENS ANTHROPIC TOOL CALLING AND GIT CREDENTIALS AS OPENHANDS SHIPS 1.10.0

LangChain fixed a critical credential leak in evaluation runs while Anthropic tool schemas get proper fallback handling, as three major agent frameworks push releases across the ecosystem.

ai-agents 54 shipped

Local LLMs Aug 6

OLLAMA FIXES KV CACHE CORRUPTION, LLAMA.CPP HARDENS FILE SEARCH, VLLM KILLS MOE LEGACY CODE

Ollama patched a critical cache corruption bug that corrupts lazy snapshots during multi-request prefills, while llama.cpp hardened Windows junction traversal and vLLM removed 14 hardcoded MoE shards across quantization methods.