{ "captured_at": "2026-07-16", "measured_against": "https://api.moonshot.ai/v1/chat/completions (Moonshot direct)", "important": "NOT measured through EvoLink. Publish gate (red line 1) requires an EvoLink call and is NOT cleared by this dataset.", "model": "kimi-k3", "official_rates_used_for_cost_math": { "input_miss_per_1m_usd": 3.0, "input_hit_per_1m_usd": 0.3, "output_per_1m_usd": 15.0, "source": "platform.kimi.ai/docs/pricing/chat-k3", "read_on": "2026-07-16" }, "runs": { "basic": { "status": 200, "elapsed_s": 3.878, "usage": { "prompt_tokens": 90, "completion_tokens": 47, "total_tokens": 137, "cached_tokens": 90, "completion_tokens_details": { "reasoning_tokens": 31 }, "prompt_tokens_details": { "cached_tokens": 90 } }, "body_keys": [ "id", "object", "created", "model", "choices", "usage" ], "message_keys": [ "role", "content", "reasoning_content" ] }, "cache_unique_prefix": { "prefix_chars": 31864, "cold": { "usage": { "prompt_tokens": 12504, "completion_tokens": 32, "total_tokens": 12536, "completion_tokens_details": { "reasoning_tokens": 16 } }, "elapsed_s": 5.117 }, "warm": { "usage": { "prompt_tokens": 12504, "completion_tokens": 32, "total_tokens": 12536, "cached_tokens": 12288, "completion_tokens_details": { "reasoning_tokens": 21 }, "prompt_tokens_details": { "cached_tokens": 12288 } }, "elapsed_s": 4.448 } }, "cache_with_key": { "cache_key_used": true, "first": { "usage": { "prompt_tokens": 12504, "completion_tokens": 32, "total_tokens": 12536, "cached_tokens": 12288, "completion_tokens_details": { "reasoning_tokens": 29 }, "prompt_tokens_details": { "cached_tokens": 12288 } }, "elapsed_s": 5.323 }, "second": { "usage": { "prompt_tokens": 12504, "completion_tokens": 32, "total_tokens": 12536, "cached_tokens": 12288, "completion_tokens_details": { "reasoning_tokens": 29 }, "prompt_tokens_details": { "cached_tokens": 12288 } }, "elapsed_s": 4.405 } }, "reasoning_overhead": { "trivial": { "usage": { "prompt_tokens": 98, "completion_tokens": 48, "total_tokens": 146, "completion_tokens_details": { "reasoning_tokens": 32 } }, "reasoning_tokens": 32, "completion_tokens": 48, "reasoning_share_pct": 66.7, "elapsed_s": 3.529 }, "moderate": { "usage": { "prompt_tokens": 96, "completion_tokens": 930, "total_tokens": 1026, "completion_tokens_details": { "reasoning_tokens": 715 } }, "reasoning_tokens": 715, "completion_tokens": 930, "reasoning_share_pct": 76.9, "elapsed_s": 27.949 }, "hard": { "usage": { "prompt_tokens": 130, "completion_tokens": 518, "total_tokens": 648, "completion_tokens_details": { "reasoning_tokens": 220 } }, "reasoning_tokens": 220, "completion_tokens": 518, "reasoning_share_pct": 42.5, "elapsed_s": 13.688 } }, "latency_short_context": { "n": 5, "ttft_s": [ 2.743, 2.805, 3.346, 3.281, 2.816 ], "ttft_median_s": 2.816, "ttft_min_s": 2.743, "ttft_max_s": 3.346, "ttfc_s": [ 4.617, 4.124 ], "total_s": [ 4.617, 5.781, 4.368, 6.261, 5.82 ], "note": "ttft = first token of any kind (reasoning or content); ttfc = first content token" }, "errors": { "401_bad_key": { "status": 401, "body": { "error": { "message": "Invalid Authentication", "type": "invalid_authentication_error" } } }, "404_wrong_model_name": { "status": 404, "body": { "error": { "message": "Not found the model kimi-k3-chat or Permission denied", "type": "resource_not_found_error" } } }, "invalid_reasoning_effort": { "status": 200, "body": { "id": "chatcmpl-6a59310458c95e6590baec90", "object": "chat.completion", "created": 1784230148, "model": "kimi-k3", "choices": [ { "index": 0, "message": { "role": "assistant", "content": "Hi there! How can I help you today?", "reasoning_content": "The user just said \"hi\" - a very simple greeting. Let me think about what the appropriate response is here.\n\nThis is a casual, friendly greeting with no question or task. The appropriate response is warm but brief - matching the energy of their message. I shouldn't dump a bunch of information or capabilities on them. A simple, friendly greeting back with perhaps a light invitation to share what they'd like help with is appropriate.\n\nFormat considerations: This should be very short - one or two sentences. No headers, no lists, no markdown. Just a conversational reply. Something like \"Hi! How can I help you today?\" or a slight variation to feel natural.\n\nI shouldn't be overly effusive or use lots of exclamation points. Just warm and natural." }, "finish_reason": "stop" } ], "usage": { "prompt_tokens": 86, "completion_tokens": 178, "total_tokens": 264, "cached_tokens": 86, "completion_tokens_details": { "reasoning_tokens": 153 }, "prompt_tokens_details": { "cached_tokens": 86 } } } }, "empty_messages": { "status": 400, "body": { "error": { "message": "Invalid request: messages must not be empty", "type": "invalid_request_error" } } }, "over_max_tokens": { "status": 200, "body": { "id": "chatcmpl-6a59310d07ba9916a203df52", "object": "chat.completion", "created": 1784230157, "model": "kimi-k3", "choices": [ { "index": 0, "message": { "role": "assistant", "content": "Hi there! How can I help you today?", "reasoning_content": "The user just said \"hi\" - a simple greeting. This is about as basic as it gets. There's no complex question, no task, no problem to solve. Just a friendly greeting.\n\nWhat's the appropriate response here? A simple, warm greeting back. Maybe an offer to help. This should be short - there's no reason to write paragraphs in response to \"hi\". A conversational, natural response is best.\n\nI shouldn't overthink this. A brief, friendly hello and perhaps a light invitation to share what they need is perfect. No headers, no bullet points, no lengthy explanation of what I can do - just a natural conversational reply." }, "finish_reason": "stop" } ], "usage": { "prompt_tokens": 86, "completion_tokens": 155, "total_tokens": 241, "cached_tokens": 86, "completion_tokens_details": { "reasoning_tokens": 130 }, "prompt_tokens_details": { "cached_tokens": 86 } } } } }, "starved_completion": { "budget_64": { "max_completion_tokens": 64, "finish_reason": "length", "reasoning_tokens": 61, "completion_tokens": 64, "content_len": 0, "content_empty": true, "billed_completion_tokens": 64 }, "budget_128": { "max_completion_tokens": 128, "finish_reason": "length", "reasoning_tokens": 125, "completion_tokens": 128, "content_len": 0, "content_empty": true, "billed_completion_tokens": 128 }, "budget_256": { "max_completion_tokens": 256, "finish_reason": "length", "reasoning_tokens": 253, "completion_tokens": 256, "content_len": 0, "content_empty": true, "billed_completion_tokens": 256 }, "budget_512": { "max_completion_tokens": 512, "finish_reason": "length", "reasoning_tokens": 509, "completion_tokens": 512, "content_len": 0, "content_empty": true, "billed_completion_tokens": 512 }, "budget_1024": { "max_completion_tokens": 1024, "finish_reason": "length", "reasoning_tokens": 1021, "completion_tokens": 1024, "content_len": 0, "content_empty": true, "billed_completion_tokens": 1024 } }, "cache_granularity": { "sections_200": { "sections": 200, "prompt_tokens": 5061, "cold_cached": 0, "warm_cached": 4864, "uncached_remainder": 197, "warm_cached_div_1024": 4.75, "cold_latency_s": 3.438, "warm_latency_s": 3.467 }, "sections_400": { "sections": 400, "prompt_tokens": 10095, "cold_cached": 0, "warm_cached": 9984, "uncached_remainder": 111, "warm_cached_div_1024": 9.75, "cold_latency_s": 3.614, "warm_latency_s": 3.775 }, "sections_800": { "sections": 800, "prompt_tokens": 20056, "cold_cached": 0, "warm_cached": 19968, "uncached_remainder": 88, "warm_cached_div_1024": 19.5, "cold_latency_s": 4.433, "warm_latency_s": 4.064 } }, "cache_key_clean": { "note": "全新前缀,首发即带 prompt_cache_key。与 C 组(不带 key)对比看是否有差异", "first_cached": 0, "second_cached": 9984, "prompt_tokens": 10080, "first_latency_s": 3.676, "second_latency_s": 4.583 }, "granularity_v2": { "n50": { "prompt_tokens": 1324, "cold_cached": 0, "warm_cached": 1280, "remainder": 44, "is_mult_256": true, "wc_div_256": 5.0, "cold_s": 3.564, "warm_s": 3.797 }, "n100": { "prompt_tokens": 2511, "cold_cached": 0, "warm_cached": 2304, "remainder": 207, "is_mult_256": true, "wc_div_256": 9.0, "cold_s": 3.018, "warm_s": 4.286 }, "n300": { "prompt_tokens": 7228, "cold_cached": 0, "warm_cached": 7168, "remainder": 60, "is_mult_256": true, "wc_div_256": 28.0, "cold_s": 3.991, "warm_s": 3.537 }, "n600": { "prompt_tokens": 14428, "cold_cached": 0, "warm_cached": 14336, "remainder": 92, "is_mult_256": true, "wc_div_256": 56.0, "cold_s": 4.381, "warm_s": 4.268 }, "n1200": { "prompt_tokens": 28986, "cold_cached": 0, "warm_cached": 28928, "remainder": 58, "is_mult_256": true, "wc_div_256": 113.0, "cold_s": 5.216, "warm_s": 3.779 } }, "ttl_probe": { "after_30s_cumulative": { "cached_tokens": 12032, "prompt_tokens": 12124, "cache_alive": true }, "after_60s_cumulative": { "cached_tokens": 12032, "prompt_tokens": 12124, "cache_alive": true }, "after_120s_cumulative": { "cached_tokens": 12032, "prompt_tokens": 12124, "cache_alive": true } }, "starve_threshold": { "budget_2048": { "finish_reason": "stop", "reasoning_tokens": 1308, "completion_tokens": 1721, "content_len": 1954, "content_empty": false } }, "long_context_latency": { "sections_4000": { "prompt_tokens": 98625, "elapsed_s": 10.495, "cost_input_usd_miss": 0.29587 }, "sections_20000": { "prompt_tokens": 497718, "elapsed_s": 52.05, "cost_input_usd_miss": 1.49315 } } } }