Chrome Achieves Dual Record-Breaking Scores on Speedometer 3.1 and JetStream 3
Key Takeaways
- ▸Chrome achieved record JetStream 3 score of 469 (10% improvement) and Speedometer 3.1 score of 61 (5% improvement)
- ▸JavaScript fast-path inlining for async operations and improved optimization heuristics delivered major performance gains
- ▸WebAssembly and SIMD optimizations enhance performance for AI, cryptography, and compute-intensive workloads
Summary
Chrome has achieved record-breaking performance on two major web benchmarks, demonstrating significant advances in browser optimization across JavaScript, WebAssembly, and compiler infrastructure. The browser improved its JetStream 3 score to 469 (a 10% gain since the start of 2026) and Speedometer 3.1 score to 61 (a 5% improvement over the past year), with both results representing dual records across all browsers. The improvements were driven by targeted optimizations in three key areas: JavaScript execution through fast-path inlining for async operations and microtask dispatch, WebAssembly performance enhancements including SIMD optimization for AI and cryptographic workloads, and compiler improvements in code generation and memory reuse efficiency.
The JavaScript optimizations focused on reducing unnecessary execution steps by inlining fast paths for common operations, with the largest gains from async operations optimization and improved heuristics for when to optimize functions. The team also enhanced BigInt handling and refined cross-function inlining strategies. WebAssembly improvements built on previous WasmGC work with new load and null-check elimination techniques, while SIMD optimization yielded significant performance gains for AI, cryptography, and interpreter use cases. These optimizations directly translate into a faster, more responsive web experience for Chrome users on real-world applications.
- Compiler improvements in code generation and memory allocation improved both performance and sandbox security
Editorial Opinion
Chrome's continued benchmark dominance reflects Google's strategic focus on optimizing the browser stack for real-world workloads, with particular attention to emerging AI compute patterns through SIMD enhancements. The improvements demonstrate how benchmark-driven development uncovers and accelerates optimization of practical use cases that traditional profiling might miss. These results underscore the competitive intensity in browser performance and the increasing importance of AI-specific optimizations as machine learning workloads migrate to the web platform.
