Google: Gemini 2.5 Flash Lite Preview 09-2025 vs Inception: Mercury 3
Neither wins outright: Google: Gemini 2.5 Flash Lite Preview 09-2025 is the pick when you need image input, Inception: Mercury 3 when you generate long outputs in one go.
Monthly cost at list price for five common workloads. Token counts per unit are stated so you can scale them to your own traffic. A workload that does not fit a model's context window is marked rather than priced.
Workload (per month)
Google: Gemini 2.5 Flash Lite Preview 09-2025
Inception: Mercury 3
Difference
Customer-support chatbot
50,000 conversations — each a 6-turn conversation (1,500 in / 400 out tokens)
$15.50
$14.50
Inception: Mercury 3 saves $1.00
RAG search over documents
20,000 questions — each a question answered from ~8 retrieved passages (6,000 in / 500 out tokens)
$16.00
$15.50
Inception: Mercury 3 saves $0.50
Coding / tool-using agent
2,000 tasks — each a multi-step task with tool calls and file context (60,000 in / 6,000 out tokens)
$16.80
$16.20
Inception: Mercury 3 saves $0.60
Batch summarisation
10,000 documents — each a 6,000-word report summarised to a page (8,000 in / 600 out tokens)
$10.40
$10.10
Inception: Mercury 3 saves $0.30
Long-document analysis
500 documents — each a 150-page contract or codebase read in one pass (100,000 in / 2,000 out tokens)
$5.40
$5.35
Inception: Mercury 3 saves $0.05
Capabilities side by side
Capability
Google: Gemini 2.5 Flash Lite Preview 09-2025
Inception: Mercury 3
What it lets you do
Tool / function calling
✓ Yes
✓ Yes
call your APIs and run agent loops
Structured output (JSON schema)
✓ Yes
✓ Yes
return JSON that validates against your schema
Reasoning / thinking mode
✓ Yes
✓ Yes
spend extra tokens thinking before answering hard problems
Image input
✓ Yes
—
read screenshots, charts and scanned pages
Audio input
✓ Yes
—
take speech or audio directly
File / PDF input
✓ Yes
—
accept documents without your own parsing step
Image output
—
—
generate images, not just text
Built-in web search
—
—
answer from live web results
Seed (reproducible sampling)
✓ Yes
—
repeat a generation for tests and evals
Log-probabilities
—
—
read token confidence for classification and scoring
From each model's published parameters and input/output types. “—” means not advertised, not proven absent.
Price history
Google: Gemini 2.5 Flash Lite Preview 09-2025: price unchanged since we started tracking it on 2026-03-10.
Inception: Mercury 3: price unchanged since we started tracking it on 2026-10-20.
Which should you choose?
Choose Google: Gemini 2.5 Flash Lite Preview 09-2025 if…
you need image input — to read screenshots, charts and scanned pages; Inception: Mercury 3 does not advertise it
you need audio input — to take speech or audio directly; Inception: Mercury 3 does not advertise it
you need file / PDF input — to accept documents without your own parsing step; Inception: Mercury 3 does not advertise it
you need seed (reproducible sampling) — to repeat a generation for tests and evals; Inception: Mercury 3 does not advertise it
Choose Inception: Mercury 3 if…
you generate long outputs in one go — up to 131K tokens (about 98,300 words) per response
you want the newer model — released 2026-10-16, 13 months after the other
Frequently asked questions
Is Google: Gemini 2.5 Flash Lite Preview 09-2025 better than Inception: Mercury 3?
Neither wins outright: Google: Gemini 2.5 Flash Lite Preview 09-2025 is the pick when you need image input, Inception: Mercury 3 when you generate long outputs in one go. In short: Google: Gemini 2.5 Flash Lite Preview 09-2025 when you need image input, you need audio input, you need file / PDF input, you need seed (reproducible sampling); Inception: Mercury 3 when you generate long outputs in one go, you want the newer model.
Which is cheaper, Google: Gemini 2.5 Flash Lite Preview 09-2025 or Inception: Mercury 3?
Inception: Mercury 3 is cheaper: $0.16 vs $0.18 per 1M tokens at a typical 3:1 input/output mix, 7% less. For 50,000 support conversations a month that is $14.50 against $15.50.
How much do Google: Gemini 2.5 Flash Lite Preview 09-2025 and Inception: Mercury 3 cost per million tokens?
Google: Gemini 2.5 Flash Lite Preview 09-2025 costs $0.10 per 1M input tokens and $0.40 per 1M output tokens; Inception: Mercury 3 costs $0.10 input and $0.35 output. Output tokens are billed separately, so long answers shift the gap.
Which has the bigger context window, Google: Gemini 2.5 Flash Lite Preview 09-2025 or Inception: Mercury 3?
Google: Gemini 2.5 Flash Lite Preview 09-2025: 1.0M tokens, about 1,573 pages, against 1M (about 1,500 pages) for Inception: Mercury 3.
Do Google: Gemini 2.5 Flash Lite Preview 09-2025 and Inception: Mercury 3 support function calling and JSON output?
Can Google: Gemini 2.5 Flash Lite Preview 09-2025 or Inception: Mercury 3 read images?
Only Google: Gemini 2.5 Flash Lite Preview 09-2025 accepts image input; Inception: Mercury 3 is text-only for input.
Is Google: Gemini 2.5 Flash Lite Preview 09-2025 or Inception: Mercury 3 free to use?
No free variant is listed for either; both are pay-per-token.
What is the maximum output length of Google: Gemini 2.5 Flash Lite Preview 09-2025 and Inception: Mercury 3?
Google: Gemini 2.5 Flash Lite Preview 09-2025 can write up to 66K tokens (about 49,200 words) per response; Inception: Mercury 3 up to 131K (about 98,300 words).
When were Google: Gemini 2.5 Flash Lite Preview 09-2025 and Inception: Mercury 3 released?
Google: Gemini 2.5 Flash Lite Preview 09-2025 was listed on 2025-09-25; Inception: Mercury 3 on 2026-10-16.
Gemini 2.5 Flash-Lite is a lightweight reasoning model in the Gemini 2.5 family, optimized for ultra-low latency and cost efficiency. It offers improved throughput, faster token generation, and better performance across common benchmarks compared to earlier Flash models. By default, "thinking" (i.e. multi-pass reasoning) is disabled to prioritize speed, but developers can enable it via the [Reasoning API parameter](https://openrouter.ai/docs/use-cases/reasoning-tokens) to selectively trade off cost for intelligence.
Mercury 3 is a diffusion-based language model from Inception, and the lab's most capable model to date. It is suited for agentic workflows and real-time applications where latency compounds, such...