All release briefsReleasedCatalog-reported release
Inception / Mercury 2.5
Mercury 2.5: diffusion-based text generation with tools and structured output
Inception’s Mercury 2.5 supports reasoning, tool use, and structured text output through an OpenAI-compatible API.
LanguageInception API
By Privaj Scout · Brief published Sep 15, 2026 · Sources checked Sep 16, 2026
SCOUT’S TAKE
A candidate for workflows where response time matters. Test quality and latency on your own workload.
- Context
- 260,000 tokens
- Standard input / output · per 1M
- $0.20 / $0.75
- Promotional input / output · per 1M
- $0.04 / $0.15 · checked Sep 16
- API identifier
- mercury-2.5
What changed.
- 01
The provider documents a 260K context window.
- 02
Reasoning can be adjusted, with support for parallel tool calls and schema-constrained output.
- 03
Mercury Voice and Router were introduced separately as previews; they are not the same access offer.
Inception announcement ↗Before you use it.
The model page currently advertises an introductory discount. Verify checkout or provider pricing before budgeting; promotional rates can change.
What the evidence showsFeatures and prices checked with Inception. The September 8 date remains catalog-reported. Vendor speed claims are not independent Scout measurements.
ut