Last verified
80% OFF AT LAUNCHDIFFUSION LLM260K CONTEXT1,100+ TOKENS/SECON INCEPTION'S OWN API

Mercury 2.5 API Pricing

Mercury 2.5 is Inception's newest diffusion LLM, generally available on the company's own API since September 8, 2026. The list price is $0.20/M input, $0.02/M cached input and $0.75/M output, the same card the OpenRouter-only preview carried, and at launch Inception takes 80% off: $0.04 / $0.004 / $0.15 is what an account is charged today. No end date is published, so the tiles hold the list rate and the discount is shown beside it. The context window is 260K, twice Mercury 2's, and Inception quotes 1,107 tokens per second.

Input - per 1M tokens
$0.20/M
Charged now $0.04/M - 80% off list
Output - per 1M tokens
$0.75/M
Charged now $0.15/M - 80% off list
Cached input - per 1M tokens
$0.02/M
Charged now $0.004/M - 80% off -90%
Effective - agentic blend
$0.108/M
92/8 split - 82% cache
§ 01 / TERMINAL

Run the numbers.

Calculator pre-loaded with Mercury 2.5's LIST rates, the ones that apply once the launch discount ends. While 80% off runs, divide every result by five for the bill you get today. Inception has not said when the discount ends.

$ /mo
Workload split
Prompt cache hit rate
Tokens you can process
—
Words equivalent (English)
—
Effective rate
—
Open full calculator (all models · share URL · CSV) →
§ 02 / SCENARIOS

Real-world presets.

CODING SUBAGENT

Context compaction

$0.014/task
120,000 in - 10,000 out~7,252 units/$100
SEARCH AGENT

Query rewrite and rerank

$0.00382/query
30,000 in - 3,000 out~26,164 units/$100
VOICE AGENT

Live call turn

$0.00054/turn
6,000 in - 300 out~185,391 units/$100
SUPPORT

Ticket reply

$0.00117/ticket
8,000 in - 1,000 out~85,528 units/$100
§ 03 / TOKENIZER

Paste text. See tokens. See cost.

What's a token →
Estimate · inception-tokenizer-estimate · ≈3.85 chars/token Auto-counts as you type

This is a chars-per-token approximation, not a real tokenizer. Actual tokens vary by language, code density, and tool-call overhead — counts are typically ±10–20% off for English prose, more for code or non-Latin scripts. For exact billing, use the vendor's official tokenizer.

Characters 494
Words 80
Tokens (estimated) 128 tokens
Cost as input · uncached $0.00003 USD
Cost as output · uncached $0.0001 USD
Cost as cached input $0.000003 USD
§ 04 / SHELF

Up against the shelf.

All models →
Model Input /M Output /M Effective blended Context Best for
Mercury 2.5 Current $0.20 cache $0.02 $0.75 $0.108 agentic 92/8 260K Latency-bound reasoning, search and voice agents
Mercury 2 $0.25 cache $0.025 $0.75 $0.12 +11% 128K The previous generation, for existing deployments
Mercury Edit 2 $0.25 cache $0.025 $0.75 $0.12 +11% 32K Autocomplete and next-edit
Mercury 2.5 Preview $0.20 cache $0.02 $0.75 $0.108 withdrawn 260K The preview this release replaced, for the record
GPT-6 Luna $0.10 cache $0.01 $0.50 $0.0682 −37% 1.05M High-volume work on a closed frontier stack
Claude Haiku 4.5 $1.00 cache $0.10 $5.00 $0.682 +6.3× 200K High-volume support and classification
Gemini 3.8 Flash $0.75 cache $0.075 $3.75 $0.481 +4.4× · promo 1M Google's fast multimodal tier
DeepSeek V4.1 Flash $0.30 cache $0.006 $1.20 $0.15 +39% 1M Cheap 1M-context agent loops with images

Frequently asked.

What Mercury 2.5 costs with and without the launch discount, what changed from the preview and from Mercury 2, and what the speed is worth.

Q · 01 How much does Mercury 2.5 cost? +
The list price is $0.20/M input, $0.02/M cached input and $0.75/M output, $0.11/M on our 92/8 agentic blend with 82% cache hits. At launch Inception takes 80% off, so an account is charged today $0.04 / $0.004 / $0.15, about $0.022/M on the same blend. Inception prints the list rates struck through beside the discounted ones on its developer table.
Q · 02 When does the 80% discount end? +
Inception has not said. The launch post calls it the price at launch, and neither the developer table nor the models page gives a date. That is why this page's tiles and calculator hold the list rate: budget on the list and treat the discount as a saving that can stop without notice.
Q · 03 What changed from Mercury 2.5 Preview? +
Where you call it, not what it costs. The preview was listed at the same $0.20 / $0.02 / $0.75 but served only through OpenRouter and absent from Inception's developer pricing table. Mercury 2.5 is on Inception's own API under the id mercury-2.5, on v1/chat/completions, with the 80% launch discount on top. The developer table no longer lists a preview row.
Q · 04 How does it compare with Mercury 2? +
At list it is cheaper on input and cached input ($0.20 and $0.02 against $0.25 and $0.025), identical on output ($0.75), with twice the context (260K against 128K) and a higher output ceiling (65,536 against 50,000 tokens). Inception claims a 40% gain in intelligence over Mercury 2, which its models page now lists as supported for existing customers. There is little reason to start a new project on Mercury 2.
Q · 05 How fast is it, and does speed change the bill? +
Inception quotes 1,107 tokens per second on widely available NVIDIA GPUs; that is the vendor's figure, not our measurement. Billing is per token as usual, so speed does not lower the token bill. It changes what latency-bound work costs in time: Inception's own examples are search pipelines that make dozens of model calls per user request, voice agents answering in under 200 milliseconds and context compaction inside coding agents.
Q · 06 Is there a free allowance? +
Inception's documentation says every new account gets a one-time 100 million free tokens, shared across all models, with no payment details required; the launch post says the same. Its marketing models page says 10 million for a new API key. We follow the documentation, but check the console before counting on the larger figure.
Q · 07 Where can I use it? +
On Inception's own OpenAI-compatible API, through Baseten, and through OpenRouter, where a reseller sets its own margin. The rates on this page are Inception's own. The model takes text in and gives text out, with tunable reasoning, parallel tool calls and schema-aligned JSON output.
Y
Reviewed by Yaroslav Vikhariev Founder · AI//COST · Prices checked by hand against docs.inceptionlabs.ai · Last verified Sep 24, 2026
Methodology Report a correction More by Y.V.