Inception Launches Mercury 3, Outperforming GPT-6 Luna and Claude Haiku 5.5 at High Speeds
- Published
- Source
- InterestingEngineering.com
Inception, a California-based AI research company, has launched Mercury 3, a diffusion language model that claims to outperform competitors GPT-6 Luna and Claude Haiku 5.5. Mercury 3 operates at speeds of four to eight times faster than these rivals, generating over 1,000 tokens per second on NVIDIA GPUs while supporting a context window of 1 million tokens.
This advancement is significant as it demonstrates that diffusion-based language models can enhance their capabilities without sacrificing speed. Traditionally, developers faced a trade-off between model capability and response speed, but Mercury 3 reportedly delivers performance that rivals or exceeds that of OpenAI and Anthropic models while maintaining high inference speeds.
Looking ahead, Inception's Mercury 3 represents a notable evolution in language model technology, building on the company's previous models since the introduction of Mercury 1 in June 2025. No further timeline was disclosed at the time of publication.
Original report
Mercury 3 claims to outperform GPT-6 Luna, Claude Haiku 5.5 at 8 times the speed
InterestingEngineering.com · Neetika WalterThis briefing is an independently written summary based on publicly available reporting and is provided for industry information and news discovery. The original report and source publication are credited and linked where applicable. RobotToday does not claim ownership of third-party source material.
Rights concerns? Contact [email protected] with the relevant URL and details. We will review the matter and take appropriate action where warranted.
More related news
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
- 10
- 11
- 12
Related Suppliers
Inception Robotics
College Park, MD startup automating hospital pharmacy workflows with MedSort, a compact robot for returned medication processing.
Barlin Times Technology 深圳市百灵时代科技有限公司
Shenzhen-based manufacturer of slip rings, rotary joints, and rotary unions for robotics, military, and industrial applications.
Polaris Motion
Maker of high-performance multi-axis motion controllers, servo drives and a real-time motion network for precision machines.
Bühler Group
Swiss family-owned multinational (founded 1860) providing process equipment and automation for grain milling, food, materials, and optical sorting.
ST Engineering
SGX-listed Singapore group building unmanned systems (Taurus UGV, Mercury AUV, DroNet UAS) plus aerospace MRO and smart-city technology.
Elephant Robotics 大象机器人
Shenzhen cobot manufacturer founded in 2016, known for the myCobot series — world's smallest 6-axis robot arms — shipping 50,000+ units/year to 51 countries.
Human Centered Robotics Lab, The University of Texas at Austin
Luis Sentis's UT Austin lab behind the DRACO humanoids, Dreamer and the ControlIt! whole-body control framework; spun off Apptronik.
The Robot Learning Company
YC X25 startup building open-source AI-native robot arms; TRLC-DK1 bimanual dev kit enables imitation learning and teleoperation for around USD 1,200.
Zyphra Technologies Inc.
Palo Alto AI company building compact language, vision-language, and text-to-speech models including the open-source Zamba and Zonos model families.
Turiya AI
California-based multi-agent AI platform enabling manufacturers to automate engineering, quoting, and production workflows via no-code/low-code SaaS.
Cerebras Systems
Sunnyvale, California AI-compute company founded 2015; builds wafer-scale WSE-3/CS-3 chips and, since August 2026, the CS-4 accelerator.
Archimedes Drive
Rijswijk, Netherlands traction-drive maker founded 2016 as a TU Delft spinout; its zero-backlash drive holds 0.2 arcmin lost motion.
Robotics needs a service framework.
RSF defines a common language for robot service capability, lifecycle operations, certification pathways, and service-provider networks.
Join the RobotToday community on LinkedIn
Daily robotics news, in-depth analysis, conference highlights, and discussions with professionals worldwide.