Parasail

We're excited to announce that @parasail_io is deploying d-Matrix Corsair™ inference accelerators alongside @nvidia Hopper and Blackwell GPUs to power heterogeneous AI inference. By pairing NVIDIA GPUs for compute-intensive prefill with Corsair for latency-sensitive decode, Parasail can deliver up to 10x faster interactive token generation with improved cost-performance and energy efficiency. The future of AI inference is heterogeneous. Read more: d-matrix.ai/announcements/… #AI #Inference #LLM #GenAI #NVIDIA #AIInfrastructure #dMatrix @Parasail_io

We’re excited to partner with @dMatrix_AI and @nvidia to deliver up to 10x faster interactive token generation. “Parasail is proving what we’ve believed from the start: the future of inference is heterogeneous,” said @sidsheth, founder and CEO of d-Matrix. By deploying Corsair alongside their NVIDIA fleet, Parasail is embracing a heterogeneous inference approach designed to optimize each phase of the workload across multiple accelerators.” D-Matrix’s incredibly fast SRAM-based inference and NVIDIA’s willingness to collaborate and open up their system makes this all possible. Link in thread.

Last month's recall of Fable was a good reminder of the supply-chain risks in adopting closed models. If a government directive or a vendor decision can pull the model your product runs on, your inference architecture has a failure point you don't control. In his latest post, Gabriel Perácio breaks down how to build defensible AI inference infrastructure (link in thread).

Control the model, control the weights, control the value of your business...

Cold-start latency is one of the biggest bottlenecks when scaling inference. Parasail's model snapshotting saves and restores CPU and GPU process state to bring vLLM replicas online 3-5x faster than rebuilding from scratch. Learn more from @MeghanaMadhyas1 in our latest blog. Link in thread.

Most inference commits ask you to predict the future. Forecast 12 months out, pick a GPU, lock in a model, sign. Then a better model drops or your traffic shifts, and you're paying for capacity you no longer need. Or worse, you can't get capacity for growth you didn't forecast. @mikehenry8484 breaks down how we rebuilt the commit around dollars of inference instead. Any model, any hardware, switch mid-contract. Link in thread.

Industry-average GPU utilization is ~5%. You pay for the other 95% anyway, even though it produces nothing. That's the idle GPU tax. Our fix: per-token billing on dedicated hardware, backed by a fluid pool. Gabriel Perácio breaks it down (link in thread).

@Zai_org Congrats! We are excited to host this on Parasail.

Introducing GLM-5.2: Frontier Intelligence, Open Weights - Significant improvements in coding and agentic tasks - Strong long-horizon capabilities with a 1M context window - Two levels of reasoning effort: GLM-5.2 (max) pushes the limits, while GLM-5.2 (high) strikes a strong balance between performance and token efficiency - MIT-licensed open weights - Same API pricing as GLM-5.1 Tech Blog: z.ai/blog/glm-5.2 Weights: huggingface.co/zai-org/GLM-5.2 API: docs.z.ai/guides/llm/glm… Coding Plan: z.ai/subscribe Chat: chat.z.ai

GLM-5.2 is live on Parasail! It's the first open weight model to really close the gap to the closed frontier on long-horizon agentic work. On benchmarks like FrontierSWE it beats GPT-5.5 and trails only Opus 4.8. We're proud to be Day-0 partners with @Zai_org.

@km_abro @MiniMax_AI The weights are open. The GPUs, sadly, are not. We offer per-token pricing though, so it's a cheap test to spin up! Check it out on our platform now, or soon on OpenRouter.

Minimax M3 is live on Parasail, day zero. It’s the first open-weight model with frontier coding & agent capabilities, 1M context window, and native multimodal understanding. DM us to spin up a dedicated endpoint and start testing!

Picking an inference endpoint comes down to three questions: - Is your workload real-time or async? - Is your traffic consistent & predictable or spiky? - Do you need an off-catalog model, SLAs, or data isolation? Full decision framework in the blog 👇

The future of AI runs on power, not just chips. Neuralwatt is building the intelligence layer to make every watt count, and they're building it on Parasail.

@RespanAI Let's go! Congrats team on the launch.

Introducing Respan AI Gateway. The world’s first AI gateway with built-in observability, evals, and prompt optimization. Access 1,000+ models through one endpoint, with the production layer teams need beyond routing.