By No Code MBA · Facts checked
- Provider
- Inception Labs
- Announcement date
- Status at last check
- Generally available
- Model family
- Mercury
- Preceding model
- Mercury 2.5
- Capabilities
- Text
What changed
Mercury 3 is a diffusion-based language model that raises the context window to 1M tokens, up from 260K in Mercury 2.5. Inception says it has over 40 times the total parameters of Mercury 1 and runs at 1,068 tokens per second on NVIDIA GPUs. Inception reports it matches or beats GPT-6 Luna and Claude Haiku 5.5 across a range of benchmarks while running 4 to 8 times faster, and completed 50% of 50 live Mind2Web browser tasks.
Access and availability
Available through the Inception API Platform (OpenAI API compatible, with 100 million free tokens), plus OpenRouter, Baseten and Vercel. A chat interface is at chat.inceptionlabs.ai, and enterprise customers can get dedicated capacity and configurable data retention through sales.
Official access documentationPricing
$0.20 per million input tokens and $0.70 per million output tokens; cached input is 90% off. A temporary launch promotion halves this to $0.10 input and $0.35 output.
Official pricingWhat you can build
Compare Mercury 3 with your current model on a representative time-sensitive task, measuring response quality, latency, and review effort on your own inputs.
Limitations
Inception does not name the benchmarks behind its headline comparison, and most performance claims come from Inception or early-access partners. Exact parameter count, maximum output length and supported input modalities are not disclosed.