Together adds Reflection Beam to serverless inference at $0.40 per 1M input tokens
By Rob Litterst, Co-founder, PricingSaaS
Published Last updated
Together’s pricing page · Change detected
What changed
Together's serverless inference pricing table now includes Reflection Beam at $0.40 per 1M input tokens, $0.04 for cached input, and $1.20 for output. It lands in the same table as Muse Spark Open's temporary free promotion, which runs through Oct 18, and sits next to Tev1 4B Experimental.
Why it matters
For teams building on Together's serverless catalog, prices move often enough — up and down — that a model's listed rate today may not hold in a week; Reflection Beam joins a table where Muse Spark Open's temporary free pricing is already running on a countdown to October 18. Watch whether Reflection Beam gets a promo of its own, or whether Together moves on to its next catalog change at its usual roughly 8-day pace.
In context
Reflection Beam lands one day after Together made Muse Spark Open free until Oct 18 and added Nemotron 3.5 Super VL — Together has made 36 real pricing changes since it started being tracked in December 2025, averaging one every 8 days. Its recent run swings in both directions: Qwen3.7-Max inference rose 25% to $2.50/$7.50 per 1M tokens in September, then Kimi K3 token prices were cut 10% weeks later even as Qwen3.8 Flash rose. Market-wide, 201 pricing changes across 153 companies have touched credits-and-usage billing in the last 180 days, 11.6% of everything tracked.