Mercury 3
Inception · Released Oct 2026
Mercury 3 is Inception's diffusion LLM with a 1M-token context at a list price of $0.20 input and $0.70 output per 1M tokens, and speed is the reason to pick it. The deciding caveat is that its quality claims come from Inception and its partners, and the headline comparisons name no benchmark. Suits voice, chat, search and browser agents where latency sets the experience, not hard reasoning.
Is it right for you?
Good for
- Latency-bound apps: Inception reports 1,068 tokens per second on NVIDIA GPUs, and says Mercury 3 targets voice and chat support, real-time tutoring and search
- Browser and web agents: TinyFish AI, an Inception partner, reports Mercury 3 completed 50% of 50 live Mind2Web navigation tasks, more than GPT-6.1-Sol (partner claim)
- Long inputs and outputs on a budget: 1M context, up to 131,072 output tokens, tool calling and structured outputs, OpenAI API compatible
Not good for
- Claims you need verified: 'outperforms GPT-6 Luna and Claude Haiku 5.5' is stated with no benchmark name, and the pages we read show no independent run
How it performs by task
High-throughput generation
1,068 tokens per second on NVIDIA GPUs per Inception (vendor claim)
Browser agents
50% of 50 live Mind2Web tasks in a TinyFish AI partner run, quoted in Inception's launch post
Pricing
Input
$0.20 / 1M tokens
Output
$0.70 / 1M tokens
Context
1M ctx; list rate shown; 50% off at launch ($0.10 / $0.35), no end date stated, so plan on the list rate
Benchmarks
No verdict changes yet
The clock starts day one. Changes land here as our verdict evolves.
Sources
- Inception: Introducing Mercury 3Oct 2026
- Inception: modelsOct 2026
Verification log
No verification checks yet
We haven't logged a verification check for this entry. Once a check runs, its history shows here.