Inception Launches Mercury 3, a Diffusion LLM Claiming 1,068 Tokens Per Second

Inception has released Mercury 3, which it describes as the most capable diffusion-based large language model to date. The company says the model outperforms GPT-6 Luna and Claude Haiku 5.5 across a range of benchmarks while running 4 to 8 times faster at more than 1,000 tokens per second.

Performance and Pricing

Mercury 3 reaches 1,068 tokens per second on NVIDIA GPUs and supports a context window of 1 million tokens, up from 260,000 for Mercury 2.5. The model is more than 40 times larger than Mercury 1, and Inception says intelligence across standard benchmarks has risen 3x from Mercury 1 mini to Mercury 3, suggesting that the scaling laws known from autoregressive models also hold for diffusion.

Pricing is set at $0.20 per million input tokens and $0.70 per million output tokens, with a launch discount of 50 percent bringing those figures to $0.10 and $0.35. Cached tokens are 90 percent off the input price.

Production Use Cases

Early-access partners are using Mercury 3 for browser agents, interactive education, real-time search and customer support. In a production web-agent harness on 50 live Mind2Web navigation tasks, Mercury 3 completed 50 percent of them, which Inception says is significantly more than GPT-6.1-Sol, using the same token volume and the same number of agent steps at a fraction of the per-token cost.

In education, General Learning, also known as RevisionDojo, reports that hints, marked answers and definitions now come back in about a second. Octen AI measured a median generation time of 0.9 seconds and a median end-to-end response time of 2.2 seconds in its Sense API tests. Lorikeet says Mercury 3 beat other models on latency while also performing higher on evals and in production for regulated customer support.

Availability and Product Portfolio

Mercury 3 is available today on the Inception API Platform and through OpenRouter, Baseten and Vercel. It is OpenAI API compatible, and enterprise deployments support dedicated capacity, autoscaling, compliance controls and configurable data retention. The company is offering 100 million free tokens for API trials.

The release expands a real-time product portfolio that includes Mercury Voice, a model customized for the low-latency requirements of voice applications that is already powering agents at Audivi AI, Altur and OpenCall, and Mercury Decide, a decision model for high-volume classification that Inception says outperforms Jev on benchmarks including JevBench and LegalBench.

Sources: Inception