Recursive: Flint (free)

{"about":"### A model for working with code\n\nFlint is a Recursive text model built for coding agents: exploring codebases, fixing bugs, implementing features and writing tests. It is an open-weight MoE model: approximately 2.5 billion of its 39.8 billion parameters are active when processing a token. The expert structure is part of its architecture; it does not mean there are several separate assistants in the chat.\n\nEureka Machine, Recursive’s research system, participated in Flint’s development. It helped explore architectural choices, select data and training methods, and optimize computation. Flint itself underwent pre-training, mid-training, instruction-example training and reinforcement learning. Its training tasks included text, code, instruction following and agentic workflows.\n\n### Understanding a project and preparing changes\n\nStart a working task with a specific part of the project: provide the necessary files as text, explain the functions’ purpose and describe the expected result. To investigate a bug, include the observed behavior and a log excerpt. Then ask for a cause explanation, a minimal fix and a test that distinguishes correct behavior from the bug. This sequence makes the result easier to verify later in your environment.\n\nFor a new feature, specify the input data, constraints and example outputs. For refactoring, list the properties that must be preserved and ask for interface changes to be described separately. The developer recommends breaking complex work into small subtasks and giving explicit instructions. A sequential conversation with intermediate results therefore fits Flint’s intended use better than an open-ended instruction to remake an entire project.\n\n### Context and connection formats\n\nIn Ailoot API, Flint accepts text and returns text. Its context window is 262,144 tokens; source material, conversation history and the response share this space. The stated output limit is 235,929 tokens. These values describe the connection’s boundaries; set the actual length to suit the task, whether a bug explanation, a patch, a test suite or a work plan.\n\nMessages, Responses and Chat Completions formats are available, allowing you to use your client application’s familiar structure. The application itself handles reading local files and applying changes: in a text request, the model receives supplied code and returns proposals, explanations or new code.\n\nFlint suits iterative technical discussions in which each step builds on the project’s contents and the verification result of the preceding change. Its specialization helps frame the task, while the usable result is concrete text: an explanation, an implementation or a test that can be run and checked.","alias_of":null,"aliases":[],"author":"Recursive","canonical_model":"recursive/flint-v1.0:free","ctx":262144,"formats":["chat","messages","responses"],"id":"recursive/flint-v1.0:free","input":["text"],"kind":"text","max_out":235929,"methods":[{"cache_read":null,"cache_write":null,"cache_write_1h":null,"cache_write_5m":null,"conditional_meter":null,"ctx":262144,"family":"anthropic","fast_multiplier":1,"flex_multiplier":1,"fmt":"messages","http":"POST","img_in":null,"img_out":null,"in":0,"input":["text"],"label":"Anthropic Messages","max_out":235929,"max_tool_calls":null,"out":0,"output":["text"],"path":"/v1/messages","request_model_id":"recursive/flint-v1.0:free","route_label":"Ultra","status_reads":[],"tiers":[],"tool_prices":null,"unit":"token","units":[]},{"cache_read":null,"cache_write":null,"cache_write_1h":null,"cache_write_5m":null,"conditional_meter":null,"ctx":262144,"family":"openai","fast_multiplier":1,"flex_multiplier":1,"fmt":"responses","http":"POST","img_in":null,"img_out":null,"in":0,"input":["text"],"label":"OpenAI Responses","max_out":235929,"max_tool_calls":null,"out":0,"output":["text"],"path":"/v1/responses","request_model_id":"recursive/flint-v1.0:free","route_label":"Ultra","status_reads":[],"tiers":[],"tool_prices":null,"unit":"token","units":[]},{"cache_read":null,"cache_write":null,"cache_write_1h":null,"cache_write_5m":null,"conditional_meter":null,"ctx":262144,"family":"openai","fast_multiplier":1,"flex_multiplier":1,"fmt":"chat","http":"POST","img_in":null,"img_out":null,"in":0,"input":["text"],"label":"OpenAI Chat","max_out":235929,"max_tool_calls":null,"out":0,"output":["text"],"path":"/v1/chat/completions","request_model_id":"recursive/flint-v1.0:free","route_label":"Ultra","status_reads":[],"tiers":[],"tool_prices":null,"unit":"token","units":[]}],"model_aliases":[{"canonical_model_id":"recursive/flint-v1.0:free","request_model_id":"recursive/flint-v1.0:free","route_label":"Ultra"}],"name":"Recursive: Flint (free)","price":{"in":0,"in_label":null,"out":0,"out_label":null,"sort_in":0,"sort_out":0,"varies":false},"public_id":"recursive/flint-v1.0:free","short":"Recursive’s text MoE model for exploring codebases, fixing bugs and preparing features and tests."}