Inception's Mercury 3 claims faster performance and benchmark wins over GPT-6 Luna and Claude Haiku 5.5

What happened

Inception released Mercury 3, a diffusion language model it says runs four to eight times faster than GPT-6 Luna and Claude Haiku 5.5 while winning eight of 12 benchmarks in its own testing. Claude leads on longer coding tasks and long-text retrieval, and no independent tests exist yet. Mercury 3 supports up to 1 million tokens, with regular pricing of $0.20 per million input tokens and $0.70 per million output tokens, discounted by half at launch.

The story so far

Inception released Mercury 3, a diffusion language model available through its own platform, OpenRouter, Baseten, and Vercel, with support for up to 1 million tokens. Inception says it runs four to eight times faster than GPT-6 Luna and Claude Haiku 5.5 and wins eight of 12 benchmarks in the company’s own testing. Claude leads on longer coding tasks and long-text retrieval. Regular pricing is $0.20 per million input tokens and $0.70 per million output tokens. Inception is offering a 50% discount at launch.

Keep reading

Trending now

1. 1Mistral and NVIDIA Announce Industrial Engineering AI Collaboration and Open Benchmarks
2. 2Mistral Releases Voxtral 3 Speech Models, Voice Agents API and LiveKit-Based SDK
3. 3PHP releases security updates after OpenAI says its models helped researchers find critical flaws
4. 4DeepSeek Open-Sources DeepInfer and API Components, With Model and Hardware Limits