| {"messages": [{"role": "system", "content": "You are a query complexity router for Mistral AI models. Given a user query, classify it into the appropriate model tier and provide a confidence score.\n\nOutput ONLY valid JSON in this exact format:\n{\"model_tier\": <1|2|3|4>, \"confidence\": <0.0-1.0>}\n\nTier definitions:\n- Tier 1 (small): Simple factual lookups, greetings, trivial math, yes/no questions\n- Tier 2 (medium): Summarization, translation, moderate reasoning, simple coding\n- Tier 3 (large): Complex analysis, multi-step reasoning, debugging, comparisons\n- Tier 4 (xlarge): Expert-level research, novel problem-solving, proofs, system design"}, {"role": "user", "content": "Write a function to find the maximum of two numbers."}, {"role": "assistant", "content": "{\"model_tier\": 1, \"confidence\": 0.92}"}]} | |
| {"messages": [{"role": "system", "content": "You are a query complexity router for Mistral AI models. Given a user query, classify it into the appropriate model tier and provide a confidence score.\n\nOutput ONLY valid JSON in this exact format:\n{\"model_tier\": <1|2|3|4>, \"confidence\": <0.0-1.0>}\n\nTier definitions:\n- Tier 1 (small): Simple factual lookups, greetings, trivial math, yes/no questions\n- Tier 2 (medium): Summarization, translation, moderate reasoning, simple coding\n- Tier 3 (large): Complex analysis, multi-step reasoning, debugging, comparisons\n- Tier 4 (xlarge): Expert-level research, novel problem-solving, proofs, system design"}, {"role": "user", "content": "Calculate 15% tip on a $45 bill."}, {"role": "assistant", "content": "{\"model_tier\": 2, \"confidence\": 0.95}"}]} | |
| {"messages": [{"role": "system", "content": "You are a query complexity router for Mistral AI models. Given a user query, classify it into the appropriate model tier and provide a confidence score.\n\nOutput ONLY valid JSON in this exact format:\n{\"model_tier\": <1|2|3|4>, \"confidence\": <0.0-1.0>}\n\nTier definitions:\n- Tier 1 (small): Simple factual lookups, greetings, trivial math, yes/no questions\n- Tier 2 (medium): Summarization, translation, moderate reasoning, simple coding\n- Tier 3 (large): Complex analysis, multi-step reasoning, debugging, comparisons\n- Tier 4 (xlarge): Expert-level research, novel problem-solving, proofs, system design"}, {"role": "user", "content": "Solve the differential equation: dy/dx = y/x with the initial condition y(1) = 2."}, {"role": "assistant", "content": "{\"model_tier\": 4, \"confidence\": 0.91}"}]} | |
| {"messages": [{"role": "system", "content": "You are a query complexity router for Mistral AI models. Given a user query, classify it into the appropriate model tier and provide a confidence score.\n\nOutput ONLY valid JSON in this exact format:\n{\"model_tier\": <1|2|3|4>, \"confidence\": <0.0-1.0>}\n\nTier definitions:\n- Tier 1 (small): Simple factual lookups, greetings, trivial math, yes/no questions\n- Tier 2 (medium): Summarization, translation, moderate reasoning, simple coding\n- Tier 3 (large): Complex analysis, multi-step reasoning, debugging, comparisons\n- Tier 4 (xlarge): Expert-level research, novel problem-solving, proofs, system design"}, {"role": "user", "content": "Write a Python server script using asyncio that handles multiple concurrent TCP connections."}, {"role": "assistant", "content": "{\"model_tier\": 4, \"confidence\": 0.89}"}]} | |
| {"messages": [{"role": "system", "content": "You are a query complexity router for Mistral AI models. Given a user query, classify it into the appropriate model tier and provide a confidence score.\n\nOutput ONLY valid JSON in this exact format:\n{\"model_tier\": <1|2|3|4>, \"confidence\": <0.0-1.0>}\n\nTier definitions:\n- Tier 1 (small): Simple factual lookups, greetings, trivial math, yes/no questions\n- Tier 2 (medium): Summarization, translation, moderate reasoning, simple coding\n- Tier 3 (large): Complex analysis, multi-step reasoning, debugging, comparisons\n- Tier 4 (xlarge): Expert-level research, novel problem-solving, proofs, system design"}, {"role": "user", "content": "Given the vector u = (1, 2, 3) and v = (4, -1, 0), calculate their cross product."}, {"role": "assistant", "content": "{\"model_tier\": 3, \"confidence\": 0.93}"}]} | |
| {"messages": [{"role": "system", "content": "You are a query complexity router for Mistral AI models. Given a user query, classify it into the appropriate model tier and provide a confidence score.\n\nOutput ONLY valid JSON in this exact format:\n{\"model_tier\": <1|2|3|4>, \"confidence\": <0.0-1.0>}\n\nTier definitions:\n- Tier 1 (small): Simple factual lookups, greetings, trivial math, yes/no questions\n- Tier 2 (medium): Summarization, translation, moderate reasoning, simple coding\n- Tier 3 (large): Complex analysis, multi-step reasoning, debugging, comparisons\n- Tier 4 (xlarge): Expert-level research, novel problem-solving, proofs, system design"}, {"role": "user", "content": "Explain the mechanism of CRISPR-Cas9 gene editing at a molecular level."}, {"role": "assistant", "content": "{\"model_tier\": 4, \"confidence\": 0.98}"}]} | |
| {"messages": [{"role": "system", "content": "You are a query complexity router for Mistral AI models. Given a user query, classify it into the appropriate model tier and provide a confidence score.\n\nOutput ONLY valid JSON in this exact format:\n{\"model_tier\": <1|2|3|4>, \"confidence\": <0.0-1.0>}\n\nTier definitions:\n- Tier 1 (small): Simple factual lookups, greetings, trivial math, yes/no questions\n- Tier 2 (medium): Summarization, translation, moderate reasoning, simple coding\n- Tier 3 (large): Complex analysis, multi-step reasoning, debugging, comparisons\n- Tier 4 (xlarge): Expert-level research, novel problem-solving, proofs, system design"}, {"role": "user", "content": "Explain the mathematical proof for the irrationality of the square root of 2."}, {"role": "assistant", "content": "{\"model_tier\": 4, \"confidence\": 0.97}"}]} | |
| {"messages": [{"role": "system", "content": "You are a query complexity router for Mistral AI models. Given a user query, classify it into the appropriate model tier and provide a confidence score.\n\nOutput ONLY valid JSON in this exact format:\n{\"model_tier\": <1|2|3|4>, \"confidence\": <0.0-1.0>}\n\nTier definitions:\n- Tier 1 (small): Simple factual lookups, greetings, trivial math, yes/no questions\n- Tier 2 (medium): Summarization, translation, moderate reasoning, simple coding\n- Tier 3 (large): Complex analysis, multi-step reasoning, debugging, comparisons\n- Tier 4 (xlarge): Expert-level research, novel problem-solving, proofs, system design"}, {"role": "user", "content": "Write a function that removes duplicate elements from a list while maintaining order."}, {"role": "assistant", "content": "{\"model_tier\": 2, \"confidence\": 0.96}"}]} | |
| {"messages": [{"role": "system", "content": "You are a query complexity router for Mistral AI models. Given a user query, classify it into the appropriate model tier and provide a confidence score.\n\nOutput ONLY valid JSON in this exact format:\n{\"model_tier\": <1|2|3|4>, \"confidence\": <0.0-1.0>}\n\nTier definitions:\n- Tier 1 (small): Simple factual lookups, greetings, trivial math, yes/no questions\n- Tier 2 (medium): Summarization, translation, moderate reasoning, simple coding\n- Tier 3 (large): Complex analysis, multi-step reasoning, debugging, comparisons\n- Tier 4 (xlarge): Expert-level research, novel problem-solving, proofs, system design"}, {"role": "user", "content": "What is the absolute value of -15?"}, {"role": "assistant", "content": "{\"model_tier\": 1, \"confidence\": 0.96}"}]} | |
| {"messages": [{"role": "system", "content": "You are a query complexity router for Mistral AI models. Given a user query, classify it into the appropriate model tier and provide a confidence score.\n\nOutput ONLY valid JSON in this exact format:\n{\"model_tier\": <1|2|3|4>, \"confidence\": <0.0-1.0>}\n\nTier definitions:\n- Tier 1 (small): Simple factual lookups, greetings, trivial math, yes/no questions\n- Tier 2 (medium): Summarization, translation, moderate reasoning, simple coding\n- Tier 3 (large): Complex analysis, multi-step reasoning, debugging, comparisons\n- Tier 4 (xlarge): Expert-level research, novel problem-solving, proofs, system design"}, {"role": "user", "content": "Create a function that converts a string to uppercase."}, {"role": "assistant", "content": "{\"model_tier\": 1, \"confidence\": 0.98}"}]} | |
| {"messages": [{"role": "system", "content": "You are a query complexity router for Mistral AI models. Given a user query, classify it into the appropriate model tier and provide a confidence score.\n\nOutput ONLY valid JSON in this exact format:\n{\"model_tier\": <1|2|3|4>, \"confidence\": <0.0-1.0>}\n\nTier definitions:\n- Tier 1 (small): Simple factual lookups, greetings, trivial math, yes/no questions\n- Tier 2 (medium): Summarization, translation, moderate reasoning, simple coding\n- Tier 3 (large): Complex analysis, multi-step reasoning, debugging, comparisons\n- Tier 4 (xlarge): Expert-level research, novel problem-solving, proofs, system design"}, {"role": "user", "content": "If 3x = 12, what is x?"}, {"role": "assistant", "content": "{\"model_tier\": 1, \"confidence\": 0.91}"}]} | |
| {"messages": [{"role": "system", "content": "You are a query complexity router for Mistral AI models. Given a user query, classify it into the appropriate model tier and provide a confidence score.\n\nOutput ONLY valid JSON in this exact format:\n{\"model_tier\": <1|2|3|4>, \"confidence\": <0.0-1.0>}\n\nTier definitions:\n- Tier 1 (small): Simple factual lookups, greetings, trivial math, yes/no questions\n- Tier 2 (medium): Summarization, translation, moderate reasoning, simple coding\n- Tier 3 (large): Complex analysis, multi-step reasoning, debugging, comparisons\n- Tier 4 (xlarge): Expert-level research, novel problem-solving, proofs, system design"}, {"role": "user", "content": "Prove by mathematical induction that for all n >= 1, the sum of the first n odd positive integers is n^2."}, {"role": "assistant", "content": "{\"model_tier\": 4, \"confidence\": 0.95}"}]} | |
| {"messages": [{"role": "system", "content": "You are a query complexity router for Mistral AI models. Given a user query, classify it into the appropriate model tier and provide a confidence score.\n\nOutput ONLY valid JSON in this exact format:\n{\"model_tier\": <1|2|3|4>, \"confidence\": <0.0-1.0>}\n\nTier definitions:\n- Tier 1 (small): Simple factual lookups, greetings, trivial math, yes/no questions\n- Tier 2 (medium): Summarization, translation, moderate reasoning, simple coding\n- Tier 3 (large): Complex analysis, multi-step reasoning, debugging, comparisons\n- Tier 4 (xlarge): Expert-level research, novel problem-solving, proofs, system design"}, {"role": "user", "content": "Write a function that returns 'Hello World'."}, {"role": "assistant", "content": "{\"model_tier\": 1, \"confidence\": 0.96}"}]} | |
| {"messages": [{"role": "system", "content": "You are a query complexity router for Mistral AI models. Given a user query, classify it into the appropriate model tier and provide a confidence score.\n\nOutput ONLY valid JSON in this exact format:\n{\"model_tier\": <1|2|3|4>, \"confidence\": <0.0-1.0>}\n\nTier definitions:\n- Tier 1 (small): Simple factual lookups, greetings, trivial math, yes/no questions\n- Tier 2 (medium): Summarization, translation, moderate reasoning, simple coding\n- Tier 3 (large): Complex analysis, multi-step reasoning, debugging, comparisons\n- Tier 4 (xlarge): Expert-level research, novel problem-solving, proofs, system design"}, {"role": "user", "content": "Does a triangle have 3 sides?"}, {"role": "assistant", "content": "{\"model_tier\": 1, \"confidence\": 0.97}"}]} | |
| {"messages": [{"role": "system", "content": "You are a query complexity router for Mistral AI models. Given a user query, classify it into the appropriate model tier and provide a confidence score.\n\nOutput ONLY valid JSON in this exact format:\n{\"model_tier\": <1|2|3|4>, \"confidence\": <0.0-1.0>}\n\nTier definitions:\n- Tier 1 (small): Simple factual lookups, greetings, trivial math, yes/no questions\n- Tier 2 (medium): Summarization, translation, moderate reasoning, simple coding\n- Tier 3 (large): Complex analysis, multi-step reasoning, debugging, comparisons\n- Tier 4 (xlarge): Expert-level research, novel problem-solving, proofs, system design"}, {"role": "user", "content": "Solve the inequality: 10 - 2x > 4"}, {"role": "assistant", "content": "{\"model_tier\": 2, \"confidence\": 0.88}"}]} | |
| {"messages": [{"role": "system", "content": "You are a query complexity router for Mistral AI models. Given a user query, classify it into the appropriate model tier and provide a confidence score.\n\nOutput ONLY valid JSON in this exact format:\n{\"model_tier\": <1|2|3|4>, \"confidence\": <0.0-1.0>}\n\nTier definitions:\n- Tier 1 (small): Simple factual lookups, greetings, trivial math, yes/no questions\n- Tier 2 (medium): Summarization, translation, moderate reasoning, simple coding\n- Tier 3 (large): Complex analysis, multi-step reasoning, debugging, comparisons\n- Tier 4 (xlarge): Expert-level research, novel problem-solving, proofs, system design"}, {"role": "user", "content": "Write a function that returns True if a number is negative."}, {"role": "assistant", "content": "{\"model_tier\": 1, \"confidence\": 0.93}"}]} | |
| {"messages": [{"role": "system", "content": "You are a query complexity router for Mistral AI models. Given a user query, classify it into the appropriate model tier and provide a confidence score.\n\nOutput ONLY valid JSON in this exact format:\n{\"model_tier\": <1|2|3|4>, \"confidence\": <0.0-1.0>}\n\nTier definitions:\n- Tier 1 (small): Simple factual lookups, greetings, trivial math, yes/no questions\n- Tier 2 (medium): Summarization, translation, moderate reasoning, simple coding\n- Tier 3 (large): Complex analysis, multi-step reasoning, debugging, comparisons\n- Tier 4 (xlarge): Expert-level research, novel problem-solving, proofs, system design"}, {"role": "user", "content": "Write a function to find the minimum of two numbers."}, {"role": "assistant", "content": "{\"model_tier\": 1, \"confidence\": 0.93}"}]} | |
| {"messages": [{"role": "system", "content": "You are a query complexity router for Mistral AI models. Given a user query, classify it into the appropriate model tier and provide a confidence score.\n\nOutput ONLY valid JSON in this exact format:\n{\"model_tier\": <1|2|3|4>, \"confidence\": <0.0-1.0>}\n\nTier definitions:\n- Tier 1 (small): Simple factual lookups, greetings, trivial math, yes/no questions\n- Tier 2 (medium): Summarization, translation, moderate reasoning, simple coding\n- Tier 3 (large): Complex analysis, multi-step reasoning, debugging, comparisons\n- Tier 4 (xlarge): Expert-level research, novel problem-solving, proofs, system design"}, {"role": "user", "content": "Write a comprehensive unit test suite using pytest for a Flask authentication blueprint, including edge cases."}, {"role": "assistant", "content": "{\"model_tier\": 3, \"confidence\": 0.93}"}]} | |
| {"messages": [{"role": "system", "content": "You are a query complexity router for Mistral AI models. Given a user query, classify it into the appropriate model tier and provide a confidence score.\n\nOutput ONLY valid JSON in this exact format:\n{\"model_tier\": <1|2|3|4>, \"confidence\": <0.0-1.0>}\n\nTier definitions:\n- Tier 1 (small): Simple factual lookups, greetings, trivial math, yes/no questions\n- Tier 2 (medium): Summarization, translation, moderate reasoning, simple coding\n- Tier 3 (large): Complex analysis, multi-step reasoning, debugging, comparisons\n- Tier 4 (xlarge): Expert-level research, novel problem-solving, proofs, system design"}, {"role": "user", "content": "Design a RESTful API specification for an e-commerce checkout flow, including endpoints, verbs, requests, and responses."}, {"role": "assistant", "content": "{\"model_tier\": 3, \"confidence\": 0.9}"}]} | |
| {"messages": [{"role": "system", "content": "You are a query complexity router for Mistral AI models. Given a user query, classify it into the appropriate model tier and provide a confidence score.\n\nOutput ONLY valid JSON in this exact format:\n{\"model_tier\": <1|2|3|4>, \"confidence\": <0.0-1.0>}\n\nTier definitions:\n- Tier 1 (small): Simple factual lookups, greetings, trivial math, yes/no questions\n- Tier 2 (medium): Summarization, translation, moderate reasoning, simple coding\n- Tier 3 (large): Complex analysis, multi-step reasoning, debugging, comparisons\n- Tier 4 (xlarge): Expert-level research, novel problem-solving, proofs, system design"}, {"role": "user", "content": "Is '123' a number or a word?"}, {"role": "assistant", "content": "{\"model_tier\": 1, \"confidence\": 0.92}"}]} | |
| {"messages": [{"role": "system", "content": "You are a query complexity router for Mistral AI models. Given a user query, classify it into the appropriate model tier and provide a confidence score.\n\nOutput ONLY valid JSON in this exact format:\n{\"model_tier\": <1|2|3|4>, \"confidence\": <0.0-1.0>}\n\nTier definitions:\n- Tier 1 (small): Simple factual lookups, greetings, trivial math, yes/no questions\n- Tier 2 (medium): Summarization, translation, moderate reasoning, simple coding\n- Tier 3 (large): Complex analysis, multi-step reasoning, debugging, comparisons\n- Tier 4 (xlarge): Expert-level research, novel problem-solving, proofs, system design"}, {"role": "user", "content": "There are 24 students in a class. Half of them are girls. How many girls are there?"}, {"role": "assistant", "content": "{\"model_tier\": 1, \"confidence\": 0.93}"}]} | |
| {"messages": [{"role": "system", "content": "You are a query complexity router for Mistral AI models. Given a user query, classify it into the appropriate model tier and provide a confidence score.\n\nOutput ONLY valid JSON in this exact format:\n{\"model_tier\": <1|2|3|4>, \"confidence\": <0.0-1.0>}\n\nTier definitions:\n- Tier 1 (small): Simple factual lookups, greetings, trivial math, yes/no questions\n- Tier 2 (medium): Summarization, translation, moderate reasoning, simple coding\n- Tier 3 (large): Complex analysis, multi-step reasoning, debugging, comparisons\n- Tier 4 (xlarge): Expert-level research, novel problem-solving, proofs, system design"}, {"role": "user", "content": "Give me a step-by-step guide to boiling an egg."}, {"role": "assistant", "content": "{\"model_tier\": 2, \"confidence\": 0.96}"}]} | |
| {"messages": [{"role": "system", "content": "You are a query complexity router for Mistral AI models. Given a user query, classify it into the appropriate model tier and provide a confidence score.\n\nOutput ONLY valid JSON in this exact format:\n{\"model_tier\": <1|2|3|4>, \"confidence\": <0.0-1.0>}\n\nTier definitions:\n- Tier 1 (small): Simple factual lookups, greetings, trivial math, yes/no questions\n- Tier 2 (medium): Summarization, translation, moderate reasoning, simple coding\n- Tier 3 (large): Complex analysis, multi-step reasoning, debugging, comparisons\n- Tier 4 (xlarge): Expert-level research, novel problem-solving, proofs, system design"}, {"role": "user", "content": "If P(A) = 0.6, P(B) = 0.5, and P(A U B) = 0.8, find P(A|B)."}, {"role": "assistant", "content": "{\"model_tier\": 3, \"confidence\": 0.9}"}]} | |
| {"messages": [{"role": "system", "content": "You are a query complexity router for Mistral AI models. Given a user query, classify it into the appropriate model tier and provide a confidence score.\n\nOutput ONLY valid JSON in this exact format:\n{\"model_tier\": <1|2|3|4>, \"confidence\": <0.0-1.0>}\n\nTier definitions:\n- Tier 1 (small): Simple factual lookups, greetings, trivial math, yes/no questions\n- Tier 2 (medium): Summarization, translation, moderate reasoning, simple coding\n- Tier 3 (large): Complex analysis, multi-step reasoning, debugging, comparisons\n- Tier 4 (xlarge): Expert-level research, novel problem-solving, proofs, system design"}, {"role": "user", "content": "What are the main differences between plant and animal cells?"}, {"role": "assistant", "content": "{\"model_tier\": 3, \"confidence\": 0.98}"}]} | |
| {"messages": [{"role": "system", "content": "You are a query complexity router for Mistral AI models. Given a user query, classify it into the appropriate model tier and provide a confidence score.\n\nOutput ONLY valid JSON in this exact format:\n{\"model_tier\": <1|2|3|4>, \"confidence\": <0.0-1.0>}\n\nTier definitions:\n- Tier 1 (small): Simple factual lookups, greetings, trivial math, yes/no questions\n- Tier 2 (medium): Summarization, translation, moderate reasoning, simple coding\n- Tier 3 (large): Complex analysis, multi-step reasoning, debugging, comparisons\n- Tier 4 (xlarge): Expert-level research, novel problem-solving, proofs, system design"}, {"role": "user", "content": "Write a smart contract in Solidity for a decentralized voting system that prevents double voting and ensures anonymity."}, {"role": "assistant", "content": "{\"model_tier\": 4, \"confidence\": 0.96}"}]} | |
| {"messages": [{"role": "system", "content": "You are a query complexity router for Mistral AI models. Given a user query, classify it into the appropriate model tier and provide a confidence score.\n\nOutput ONLY valid JSON in this exact format:\n{\"model_tier\": <1|2|3|4>, \"confidence\": <0.0-1.0>}\n\nTier definitions:\n- Tier 1 (small): Simple factual lookups, greetings, trivial math, yes/no questions\n- Tier 2 (medium): Summarization, translation, moderate reasoning, simple coding\n- Tier 3 (large): Complex analysis, multi-step reasoning, debugging, comparisons\n- Tier 4 (xlarge): Expert-level research, novel problem-solving, proofs, system design"}, {"role": "user", "content": "Create a function that returns the absolute value of a number."}, {"role": "assistant", "content": "{\"model_tier\": 1, \"confidence\": 0.96}"}]} | |
| {"messages": [{"role": "system", "content": "You are a query complexity router for Mistral AI models. Given a user query, classify it into the appropriate model tier and provide a confidence score.\n\nOutput ONLY valid JSON in this exact format:\n{\"model_tier\": <1|2|3|4>, \"confidence\": <0.0-1.0>}\n\nTier definitions:\n- Tier 1 (small): Simple factual lookups, greetings, trivial math, yes/no questions\n- Tier 2 (medium): Summarization, translation, moderate reasoning, simple coding\n- Tier 3 (large): Complex analysis, multi-step reasoning, debugging, comparisons\n- Tier 4 (xlarge): Expert-level research, novel problem-solving, proofs, system design"}, {"role": "user", "content": "Create a python script to print the numbers 1 to 5."}, {"role": "assistant", "content": "{\"model_tier\": 1, \"confidence\": 0.96}"}]} | |
| {"messages": [{"role": "system", "content": "You are a query complexity router for Mistral AI models. Given a user query, classify it into the appropriate model tier and provide a confidence score.\n\nOutput ONLY valid JSON in this exact format:\n{\"model_tier\": <1|2|3|4>, \"confidence\": <0.0-1.0>}\n\nTier definitions:\n- Tier 1 (small): Simple factual lookups, greetings, trivial math, yes/no questions\n- Tier 2 (medium): Summarization, translation, moderate reasoning, simple coding\n- Tier 3 (large): Complex analysis, multi-step reasoning, debugging, comparisons\n- Tier 4 (xlarge): Expert-level research, novel problem-solving, proofs, system design"}, {"role": "user", "content": "Write a function to check if two strings are anagrams of each other."}, {"role": "assistant", "content": "{\"model_tier\": 2, \"confidence\": 0.9}"}]} | |
| {"messages": [{"role": "system", "content": "You are a query complexity router for Mistral AI models. Given a user query, classify it into the appropriate model tier and provide a confidence score.\n\nOutput ONLY valid JSON in this exact format:\n{\"model_tier\": <1|2|3|4>, \"confidence\": <0.0-1.0>}\n\nTier definitions:\n- Tier 1 (small): Simple factual lookups, greetings, trivial math, yes/no questions\n- Tier 2 (medium): Summarization, translation, moderate reasoning, simple coding\n- Tier 3 (large): Complex analysis, multi-step reasoning, debugging, comparisons\n- Tier 4 (xlarge): Expert-level research, novel problem-solving, proofs, system design"}, {"role": "user", "content": "Why do leaves change color in the fall?"}, {"role": "assistant", "content": "{\"model_tier\": 2, \"confidence\": 0.95}"}]} | |
| {"messages": [{"role": "system", "content": "You are a query complexity router for Mistral AI models. Given a user query, classify it into the appropriate model tier and provide a confidence score.\n\nOutput ONLY valid JSON in this exact format:\n{\"model_tier\": <1|2|3|4>, \"confidence\": <0.0-1.0>}\n\nTier definitions:\n- Tier 1 (small): Simple factual lookups, greetings, trivial math, yes/no questions\n- Tier 2 (medium): Summarization, translation, moderate reasoning, simple coding\n- Tier 3 (large): Complex analysis, multi-step reasoning, debugging, comparisons\n- Tier 4 (xlarge): Expert-level research, novel problem-solving, proofs, system design"}, {"role": "user", "content": "Write a short poem about a cat."}, {"role": "assistant", "content": "{\"model_tier\": 2, \"confidence\": 0.93}"}]} | |