Mercury 3
by Inception
Today's price · per 1M tokens
Input
$0.200
per 1M tokens
Output
$0.700
per 1M tokens
Blended
$0.325
blended $/1M — 3:1 weighted input:output
Cached input
$0.020
10% of input — prompt caching
How this price is scoped: These are Inception's list prices, per 1M tokens: $0.20 input / $0.02 cached input / $0.70 output. Mercury 3 launched on 2026-10-19 at 50% off, so the rate you pay today is $0.10 input / $0.01 cached input / $0.35 output. Inception's rate card shows both, the list price struck through beside the discounted one, under a "50% off Mercury 3" banner, and states no end date for the discount. We publish the list price, the rate the model returns to, rather than a discounted number that could end without notice. Verified 2026-10-20.
Price receipt
We read Inception's own pricing page on and found Mercury 3 listed with a price on the same row — that read is where the number above comes from, recorded as a new model. We keep the page text we read; its content hash is 121a84c65f.
Overview
Inception's third-generation diffusion large language model (dLLM) and its most capable model so far, succeeding Mercury 2.5. Like the rest of the Mercury family it refines a whole answer in parallel over a few denoising passes instead of decoding left to right; Inception reports more than 1,000 tokens per second on NVIDIA GPUs from a model over 40 times the size of Mercury 1. Context grows from 260K to 1M tokens, with up to 131,072 output tokens, tool calling, structured outputs and four reasoning-effort levels (instant, low, medium, high). List price $0.20 / $0.70 per 1M tokens, 50% off at launch.
Capabilities
struck through = not supportedSpecifications
- Provider
- Inception
- Context window
- 1M tokens
- Modality
- text
- Parameters
- Proprietary
- Open source
- No — proprietary
- Released
- Status
- Current
- Last updated
- Tags
Availability verified: — listed on Inception's own page