Mercury 2.5 Launches with 1,107 Tokens/sec Inference Speed
volokuleshov · x · 2026-09-02
Inception AI released Mercury 2.5, the next generation of its Mercury diffusion models, achieving 1,107 tokens/sec via parallel token generation. The update focuses on agentic workloads, featuring tunable reasoning, parallel tool calls, and schema-aligned JSON output. It is now live on OpenRouter.
Related event: Inception AI's Mercury 2.5 Hits Over 1,100 Tokens Per Second(2 posts)→
More from Models
- Mocked benchmark shows DeepSeek's document chart grounding badly misaligned — burkov · 2026-10-18
- Dev complains Claude's habit of saying "not done" is his biggest pain point — jasonkneen · 2026-10-18
- Open-weight TYGER speech decision model makes 80ms turn-taking calls for full-duplex voice agents — solyarisoftware · 2026-10-18
- Paying Google AI subscriber blocked: custom-domain Workspace accounts can't access AI Studio — ChrisRackauckas · 2026-10-18
- Running Ling 3.1 Flash (560B MoE, 25B Active) Across 4 DGX Sparks via llama.cpp RPC — BerryBlush_Rogers691 · 2026-10-18
- Users press OpenAI on long-delayed Adult Mode a year after it was promised — ___Patrice___ · 2026-10-18