Ben's Bites
·
1 month ago
Google I/O is launching today with Gemini models reportedly benchmarking similarly to GPT-5.5, though practical performance in real-world use remains uncertain. Anthropic acquired Stainless to build SDKs and added self-hosted sandboxes to Claude Managed Agents, while Cursor released Composer 2.5 trained partly on SpaceX GPUs at lower cost than competing models. The shift toward agent-first development means framework importance will decline in favor of managing agents, sandboxing, and local versus cloud execution.
Together AI
·
1 month ago
Together released a benchmark measuring inference performance on production coding agent workloads, where Together Inference Engine delivers 31% higher throughput than TensorRT-LLM and maintains 2× better time-to-first-token at saturation of 2.5M tokens per minute. The company optimized performance through ThunderMLA (a fused kernel for Multi-head Latent Attention), custom kernels, and full-stack profiling rather than relying on single-user benchmarks. For typical coding requests of 80k-100k input tokens, Kimi K2.7 on Together costs $0.029 per request compared to $0.097 for Claude Opus, saving an engineering team of 150 people approximately $421k annually.
