Product Update
Blog posts in the Product Update category
Product Update
September 29, 2026Photon 2.6: 400+ tok/s with Qwen 3.5
Photon runs Qwen3.5-27B at over 400 tokens per second on one B200. FP8 and DFlash speculative verification are now part of our megakernel compiler.
Read more
Product Update
September 24, 2026Photon can now speak
Up to 70% lower speech startup latency than vLLM-Omni. Qwen3-TTS and Kokoro are now available in Photon.
Read more