← All model drops
Live drop
Mercury 2.5
InceptionDropped Sep 8, 202632 builds22 makers14 prompts
Diffusion LLM at 1,107 tok/s with 260K context for instant agent loops and subagent workflows.
Live drop
InceptionDropped Sep 8, 202632 builds22 makers14 prompts
Diffusion LLM at 1,107 tok/s with 260K context for instant agent loops and subagent workflows.
About this release
Mercury 2.5 is Inception's latest diffusion-based reasoning LLM. It delivers a 40% intelligence gain over Mercury 2 while maintaining the same low-latency profile, reaching 1,107 tokens per second on standard NVIDIA GPUs. The model supports tunable reasoning effort, parallel tool calls, schema-aligned JSON output, and a 260K-token context window. It targets production workloads in search agents, voice pipelines, and coding subagents where latency compounds across many calls.
Read the official Inception launch pageWorth opening first
𝕏GamesGTA6: Shenzhen LiteA playable browser-based open-world game prototype of Shenzhen built with Babylon.jsby 地球人
𝕏Appsinvideo AITurn text prompts and ideas into publish-ready videos with AI.by Invideo
𝕏VideoLess Light Pollution, More LifeA cinematic AI video exploring the visual impact of light pollution on the natural world.by Jessica Collins
𝕏AppsBlyth AIA unified creative workspace for AI image, video, and 3D generation.by Yusuf K
𝕏ToysLAND OF OOOInteractive voxel reconstruction of Adventure Time's Land of Ooo built with Gemini.by Inmersion𝕏ArtUnited States Continuous Line ArtA complex cityscape of the US drawn with a single continuous line and wax-crayon accents.by simeon-sanai
𝕏agentUltracodeMulti-agent swarm orchestration for Claude Code.by Daniel Mac
𝕏VideoAI UGC 'One Character' StrategyPaste the URL and Dropday.ai will credit you, track the build as it evolves, and put it on this drop.
Submit a build
