LOCAL LLM FIELD GUIDE

Choose Local LLMs by work: Thai, chatbot, agent, coding, workflow

Do not start with a supposedly best model. Profile language and domain fit, tool calling, coding, structured output, context, memory floor, and runtime, then evaluate your workload with a disclosed test set.

WORKLOAD PROFILE

Compare candidates with one scorecard

Thai: Domain, grounding, tone
Agent: Tool schema, recovery, planning
Coding: Patch, tests, stack
Workflow: JSON schema, retry

Thai explanatory diagramThai-language diagram. The English article text provides the matching terminology and caveats.

USE IT ON REAL WORK

Use one scorecard for every candidate

A model name does not tell you whether it suits the work. Build a small evaluation set from real prompts, tools, documents, and user-visible failures.

Pass the quality bar first, then compare memory and tok/s in one runtime. That leaves an explainable selection record.

Work	Pass criterion	Failure case to include
Thai	Correct instructions, tone, and domain terms	Ungrounded answer or wrong term
chatbot	Coherent context-aware replies	Topic loss across turns
agent	Tool calls match the schema	Cannot recover from a tool error
coding	Patch passes relevant tests	Regression outside the intended change
workflow	JSON matches schema and retries	Missing field or wrong type

A minimum evaluation pack

Ten Thai prompts from real work
A multi-turn conversation with source documents
One malformed tool result
A code change with test feedback
Structured output with invalid JSON and retry

Example structured-output contract

{
  "ticket_id": "string",
  "category": "billing | support | security",
  "needs_human": true
}

Thai language

Test your Thai tasks directly: instruction following, retrieval grounding, policy tone, and domain terms. A Thai benchmark needs a named benchmark, prompt format, and exact model release. the ThaiLLM model card is a source starting point, not a substitute for workload evidence.

Chatbot and agent

Chatbots need conversational quality, safety, and long context. Agents need tool-call schema behavior, recovery after tool errors, planning, and latency. Test multi-turn and failure cases, not only a single prompt.

Coding and structured workflow

Coding evaluation should match the team language and stack, patch correctness, and test-feedback loop. Workflows need JSON/schema validity, extraction, routing, and deterministic retry. A fluent model is not necessarily the most stable structured-output model.

Memory floor and deployability

A strong model that cannot fit required quantization, context, and concurrency is not deployable. Shortlist two to four candidates that load in the runtime, then compare cost and quality with one evaluation set and the calculator.

Sources

ThaiLLM-8B model card

Try your configuration

Put the intended model, quantization, context, framework, and hardware into the calculator, then validate with a production-like benchmark.

Open calculator