Browse executable prompts and workflows by task, content type, and runtime. Each card shows inputs, verification status, and the original source.
Runtimes vary by resource and may include an API, CLI, agent framework, or browser coding environment. Check each card for details.
Editor picks
Editor picks
2
UnverifiedHugging Face
LongCat-Flash-Thinking-2601: Official Chat Template, Tool Calling, and Reasoning-History Configuration
Configure the LongCat-Flash-Thinking-2601 chat template with an explicit reasoning-history field, run one research question with a retrieval tool, and check the trace separately from the answer.
Content typeConfiguration
InputsResearch question, Reasoning history or an explicit empty history, Tool JSON schema, Retrieval result, Trace acceptance rules
StatusA service using the model tokenizer; tool schema, retrieval result, and trace logs must be capturable.
LongCat-Flash-Thinking-2601: Official SGLang/vLLM Deployment and MTP Configuration
Follow the official deployment notes to start MTP in SGLang or vLLM, hold concurrency and context constant, and measure first-token latency, generation speed, and tool-call parsing.
Content typeConfiguration
InputsBackend and version, GPU count and precision, Concurrency and context settings, Fixed test prompts, Latency and parsing logs
StatusGPU node, pinned SGLang/vLLM version, model weights, and MTP settings; a benchmark script is required.
LongCat-Flash-Thinking-2601: Official Chat Template, Tool Calling, and Reasoning-History Configuration
Configure the LongCat-Flash-Thinking-2601 chat template with an explicit reasoning-history field, run one research question with a retrieval tool, and check the trace separately from the answer.
Content typeConfiguration
InputsResearch question, Reasoning history or an explicit empty history, Tool JSON schema, Retrieval result, Trace acceptance rules
StatusA service using the model tokenizer; tool schema, retrieval result, and trace logs must be capturable.
LongCat-Flash-Thinking-2601: Official SGLang/vLLM Deployment and MTP Configuration
Follow the official deployment notes to start MTP in SGLang or vLLM, hold concurrency and context constant, and measure first-token latency, generation speed, and tool-call parsing.
Content typeConfiguration
InputsBackend and version, GPU count and precision, Concurrency and context settings, Fixed test prompts, Latency and parsing logs
StatusGPU node, pinned SGLang/vLLM version, model weights, and MTP settings; a benchmark script is required.
