Models & Labs
Together AI Enhances Model Inference Configuration
Together AI Bloghigh confidence
Why it matters
- →Enables seamless rollouts and A/B testing with zero downtime.
- →Ensures efficient resource allocation and scaling through capacity-aware traffic splitting.
- →Provides consistent performance with immutable configurations and easy rollback options.
Together AI has unveiled a new architecture for model inference that combines endpoints, deployments, and configurations with a capacity-aware traffic split. This setup allows for advanced features like rollouts and A/B testing while maintaining zero-downtime updates. The platform uses immutable configurations to ensure consistent performance and easy rollback options. This approach simplifies the deployment process and enhances the reliability of AI applications by optimizing resource allocation and scaling.
Read originalThe story around this
Earlier coverage that leads up to this article, and what followed. Lines connect each piece to the closest one after it, converging here.
How we got here
- 1JBS Dev on AI's Last Mile and Imperfect Data
AI News · May 12, 2026 · Background
- 2vLLM v0.22.0 Release Enhances Model Performance
vLLM Releases · May 31, 2026 · Related
- 3OpenAI Unveils Deployment Simulation for AI Models
OpenAI · June 16, 2026 · Related
- 4Enterprise AI Focuses on Inference Optimization
The AI Daily Brief · June 19, 2026 · Related
- 5MIT and Microsoft Enhance AI Workflow Efficiency
MIT News AI · June 25, 2026 · Related
- 6Scaling AI Agents with Redis Iris
Cole Medin · July 9, 2026 · Related
- 7Hugging Face Rethinks Model Routing as System Optimization
Hugging Face Blog · July 15, 2026 · Related
- 8Model Context Protocol Update Eases AI Integration
TechCrunch AI · July 20, 2026 · Related
What happened next
- 9AI Infrastructure Demands New Architectural Approach
MIT Technology Review AI · September 4, 2026 · Related
- 10Together AI automates model rollouts with metric gates
Together AI Blog · September 22, 2026 · Same story
- 11Jev as a Smart Router for Agent Harnesses
Sam Witteveen · September 29, 2026 · Related
- 12Liquid AI releases efficient on-device agent model
Hugging Face Blog · October 13, 2026 · Related
- 13Together AI automates inference optimization with agents
Together AI Blog · November 4, 2026 · Same story
Follow this story
Open the full story →Together AI Enhances Open-Weight Inference Platform
3 developments
- Jul 23 · Together AI Blog
Together AI Enhances Open-Weight Inference Platform - Jul 29 · Together AI Blog
Together AI Enhances Model Inference Configuration (This article)↳ Together AI adds capacity-aware traffic splitting for seamless rollouts, A/B testing, and zero-downtime updates using immutable configuratio - Jul 31 · Together AI Blog
Together AI Introduces Autoscaling for LLM Inference↳ Together AI introduces autoscaling for LLM inference based on in-flight requests and GPU utilization to address cold starts and misleading C
More in Models & Labs
Models & Labsother
llama.cpp b12340 adds CUDA 13 and ROCm 10.0
This release quietly closes the hardware gap for local inference by adding default support for CUDA 13 and ROCm 10.0 across Linux and Windows. Users with newer NVIDIA GPUs or AMD accelerators no longer need to compile from source to get performance, while Apple Silicon builds now ship KleidiAI disabled by default, shifting the focus back to standard ARM optimizations. The inclusion of Snapdragon support for Linux and Android marks a significant step toward on-device AI beyond the x86/ARM desktop ecosystem. It is less about new features and more about ensuring llama.cpp remains the universal runtime for every available accelerator.
llama.cpp Releases
Models & Labscoding
llama.cpp b12342 adds few-row MMQ for speculative decoding
Speculative decoding just got a significant speed boost on NVIDIA GPUs. This release introduces a specialized kernel for 'few-row' matrix multiplications, targeting the small batch sizes typical of verification passes in speculative decoding workflows. By optimizing these specific operations for quantized formats like Q8_0 and IQ4 variants, llama.cpp reduces latency where it matters most during token generation. The change is CUDA-only for now but leverages hardware capabilities from Ampere onwards, making local inference faster without requiring new model architectures or complex configuration changes.
llama.cpp Releases
© TechCrunch AIModels & Labsmodels
Google deploys Gemini 4 Hydrogen for faster Search
Google is quietly shifting its search infrastructure by replacing older fast models with Gemini 4 Hydrogen, a tier designed specifically for speed and cost-efficiency rather than raw reasoning power. This model completes tasks 2.5 times faster than its predecessor while beating the more powerful Argon on specific intelligence benchmarks, proving that specialized efficiency can outperform brute force in high-volume workflows. By integrating this into Search's fan-out technique, Google now breaks down complex queries into subqueries with greater precision and speed, delivering richer visual responses without the latency of heavier models. This marks a strategic pivot toward agentic execution layers within consumer search, making AI Mode feel snappier for everyday multi-part questions.
TechCrunch AI