Post-Cutoff

InceptionMercury familyCurrentReasoning modelReleased 8 Sep 2026

Mercury 2.5

Mercury 2.5 is aimed at latency-sensitive work (voice agents, autocomplete, agent sub-calls) where speed matters more than peak intelligence.

Model idmercury-2.5

Price$0.20 in, $0.75 out per 1M tokens

Checked30 September 2026

No published training cutoff

At least1,680events since its release on 8 September 2026

65 historic, 373 major. 72 days of news.

Inception doesn’t publish a training cutoff, so we count from its release date. The real gap is larger.

Five things it has never heard of

How to call it

3 ways to use Mercury 2.5, with the model id each provider expects.

Providers, model ids, endpoints and docs for Mercury 2.5
ProviderModel idEndpointDocs
Inception API (OpenAI-compatible) mercury-2.5 https://api.inceptionlabs.ai/v1/chat/completions docs.inceptionlabs.ai
OpenRouter inception/mercury-2.5 openrouter.ai
Baseten baseten.co

Pricing

Prices for Mercury 2.5
Input$0.20 per 1M tokens
Output$0.75 per 1M tokens

As published: “per 1M tokens (list price; 80% launch discount brings it to $0.04 / $0.15, end date not stated)”.

Source: docs.inceptionlabs.ai, checked 30 September 2026.

Specs

Input
Text
Output
Text
Context window
260,000 tokens
Max output
65,536 tokens
Open weights
No
Licence
Proprietary
Training cutoff
Not published
Released
8 September 2026
Status
Current
Type
Reasoning model

What stands out

What its maker marketed at launch, and what people found later, each with its source.

  • Diffusion-based text generation at very high speed

    A diffusion LLM (dLLM) that refines many tokens in parallel rather than one at a time; Inception claims 1,107 tokens/s on NVIDIA GPUs, and Artificial Analysis measured ~661 tokens/s.

    Source: inceptionlabs.ai

  • Fast reasoning with tool calling

    Configurable reasoning effort, tool calling and structured outputs; Inception claims a 40% intelligence gain over Mercury 2 and places it near GPT-5.6 Luna (low), Gemini 3.5 Flash-Lite and Claude Haiku 4.5.

    Source: inceptionlabs.ai

Notes

Artificial Analysis lists the release as Sept 8, 2026, measures ~661 tok/s and gives an Intelligence Index of 12, below average for its price tier. It lists $0.25/$0.75, while Inception’s docs list $0.20/$0.75 before the discount. Inception’s blog page showed Sept 29, 2026 when fetched, and OpenRouter also has a separate inception/mercury-2.5-preview listing, so the Sept 8 date may be the preview. Speed figures are vendor claims without disclosed batch or hardware details (RuntimeWire). Customer claims: Augment Code context compaction 150 s → 27 s; OpenCall ~170 ms median latency. 100M free API tokens for new accounts.

Sources: Inception blog · Artificial Analysis · RuntimeWire · Inception on X