{
  "generatedBy": "bench.mjs",
  "startedAt": "2026-09-26T19:36:38.403Z",
  "finishedAt": "2026-09-26T19:39:44.112Z",
  "machine": {
    "cpu": "Apple M1",
    "os": "Darwin 26.3",
    "arch": "arm64",
    "node": "v22.18.0",
    "python": "Python 3.13.5"
  },
  "metricCount": 93,
  "metrics": [
    {
      "metric": "tarball_bytes",
      "value": 160042307,
      "unit": "bytes",
      "note": "ollama-darwin.tgz release asset"
    },
    {
      "metric": "install_dir_mb",
      "value": 497,
      "unit": "MB",
      "note": "unpacked standalone binary + bundled libraries"
    },
    {
      "metric": "no_server_list_output",
      "value": "Error: could not connect to ollama server, run 'ollama serve' to start it",
      "unit": "text",
      "note": "ollama list before any server is running"
    },
    {
      "metric": "client_version_no_server",
      "value": "Warning: could not connect to a running Ollama instance Warning: client version is 0.34.4",
      "unit": "text",
      "note": "ollama --version with no server"
    },
    {
      "metric": "cli_subcommands",
      "value": "serve create show run stop pull push signin signout list ps cp rm launch help",
      "unit": "text",
      "note": "Available Commands in ollama --help"
    },
    {
      "metric": "cli_subcommand_count",
      "value": 15,
      "unit": "count",
      "note": "subcommands listed by ollama --help"
    },
    {
      "metric": "serve_env_vars",
      "value": "OLLAMA_DEBUG OLLAMA_HOST OLLAMA_CONTEXT_LENGTH OLLAMA_KEEP_ALIVE OLLAMA_MAX_LOADED_MODELS OLLAMA_MAX_TRANSFER_STREAMS OLLAMA_MAX_QUEUE OLLAMA_MODELS OLLAMA_NUM_PARALLEL OLLAMA_NO_CLOUD OLLAMA_NOPRUNE OLLAMA_ORIGINS OLLAMA_SCHED_SPREAD OLLAMA_FLASH_ATTENTION OLLAMA_KV_CACHE_TYPE OLLAMA_LLM_LIBRARY OLLAMA_GPU_OVERHEAD OLLAMA_IGPU_ENABLE LLAMA_ARG_FIT LLAMA_ARG_FIT_TARGET OLLAMA_LOAD_TIMEOUT",
      "unit": "text",
      "note": "Environment Variables in ollama serve --help"
    },
    {
      "metric": "serve_env_var_count",
      "value": 21,
      "unit": "count",
      "note": "env vars documented by ollama serve --help"
    },
    {
      "metric": "serve_env_help",
      "value": "      OLLAMA_DEBUG                  Show additional debug information (e.g. OLLAMA_DEBUG=1)\n      OLLAMA_HOST                   IP Address for the ollama server (default 127.0.0.1:11434)\n      OLLAMA_CONTEXT_LENGTH         Context length to use unless otherwise specified (default: 4k/32k/256k based on VRAM)\n      OLLAMA_KEEP_ALIVE             The duration that models stay loaded in memory (default \"5m\")\n      OLLAMA_MAX_LOADED_MODELS      Maximum number of loaded models per GPU\n      OLLAMA_MAX_TRANSFER_STREAMS   Maximum parallel transfer streams for safetensors model pulls/pushes (default 4)\n      OLLAMA_MAX_QUEUE              Maximum number of queued requests\n      OLLAMA_MODELS                 The path to the models directory\n      OLLAMA_NUM_PARALLEL           Maximum number of parallel requests\n      OLLAMA_NO_CLOUD               Disable Ollama cloud features (remote inference and web search)\n      OLLAMA_NOPRUNE                Do not prune model blobs on startup\n      OLLAMA_ORIGINS                A comma separated list of allowed origins\n      OLLAMA_SCHED_SPREAD           Always schedule model across all GPUs\n      OLLAMA_FLASH_ATTENTION        Enabled flash attention\n      OLLAMA_KV_CACHE_TYPE          Quantization type for the K/V cache (default: f16)\n      OLLAMA_LLM_LIBRARY            Set LLM library to bypass autodetection\n      OLLAMA_GPU_OVERHEAD           Reserve a portion of VRAM per GPU (bytes)\n      OLLAMA_IGPU_ENABLE            Enable integrated GPUs\n      LLAMA_ARG_FIT                 Enable llama.cpp automatic fit of unset memory options (default \"on\")\n      LLAMA_ARG_FIT_TARGET          Target free VRAM margin per device for llama.cpp fit (MiB)\n      OLLAMA_LOAD_TIMEOUT           How long to allow model loads to stall before giving up (default \"5m\")",
      "unit": "text",
      "note": "full env var help text from ollama serve --help"
    },
    {
      "metric": "run_env_help",
      "value": "      OLLAMA_EDITOR                 Path to editor for interactive prompt editing (Ctrl+G)\n      OLLAMA_HOST                   IP Address for the ollama server (default 127.0.0.1:11434)\n      OLLAMA_NOHISTORY              Do not preserve readline history",
      "unit": "text",
      "note": "ollama run --help env vars"
    },
    {
      "metric": "run_flags",
      "value": "      --dimensions int          Truncate output embeddings to specified dimension (embedding models only)\n      --format string           Response format (e.g. json)\n  -h, --help                    help for run\n      --hidethinking            Hide thinking output (if provided)\n      --insecure                Use an insecure registry\n      --keepalive string        Duration to keep a model loaded (e.g. 5m)\n      --nowordwrap              Don't wrap words to the next line automatically\n      --think string[=\"true\"]   Enable thinking mode: true/false or high/medium/low for supported models\n      --truncate                For embedding models: truncate inputs exceeding context length (default: true). Set --truncate=false to error instead\n      --verbose                 Show timings for response",
      "unit": "text",
      "note": "ollama run --help flags"
    },
    {
      "metric": "show_flags",
      "value": "  -h, --help         help for show\n      --license      Show license of a model\n      --modelfile    Show Modelfile of a model\n      --parameters   Show parameters of a model\n      --system       Show system message of a model\n      --template     Show template of a model\n  -v, --verbose      Show detailed model information",
      "unit": "text",
      "note": "ollama show --help flags"
    },
    {
      "metric": "create_flags",
      "value": "      --draft-quantize string   Quantize safetensors draft model to this level\n  -f, --file string             Name of the Modelfile (default \"Modelfile\")\n      --force                   Continue local creation when MLX validation fails\n  -h, --help                    help for create\n  -q, --quantize string         Quantize safetensors model to this level (e.g. nvfp4)",
      "unit": "text",
      "note": "ollama create --help flags"
    },
    {
      "metric": "list_aliases",
      "value": "list, ls",
      "unit": "text",
      "note": "ollama list --help aliases"
    },
    {
      "metric": "serve_aliases",
      "value": "serve, start",
      "unit": "text",
      "note": "ollama serve --help aliases"
    },
    {
      "metric": "serve_ready_ms",
      "value": 247,
      "unit": "ms",
      "note": "time from ollama serve start to first HTTP 200 on :11434, n=1"
    },
    {
      "metric": "root_endpoint_body",
      "value": "Ollama is running",
      "unit": "text",
      "note": "GET http://127.0.0.1:11434/"
    },
    {
      "metric": "api_version",
      "value": "{\"version\":\"0.34.4\"}",
      "unit": "json",
      "note": "GET /api/version"
    },
    {
      "metric": "server_version",
      "value": "0.34.4",
      "unit": "text",
      "note": "server version"
    },
    {
      "metric": "listen_line",
      "value": "Listening on 127.0.0.1:11434",
      "unit": "text",
      "note": "serve log listen line"
    },
    {
      "metric": "default_models_dir",
      "value": "OLLAMA_MODELS:~/.ollama/models",
      "unit": "text",
      "note": "OLLAMA_MODELS reported in server config log (HOME shown as ~)"
    },
    {
      "metric": "default_keep_alive",
      "value": "OLLAMA_KEEP_ALIVE:5m0s",
      "unit": "text",
      "note": "server config log"
    },
    {
      "metric": "default_num_parallel",
      "value": "OLLAMA_NUM_PARALLEL:1",
      "unit": "text",
      "note": "server config log"
    },
    {
      "metric": "default_context_length",
      "value": "OLLAMA_CONTEXT_LENGTH:0",
      "unit": "text",
      "note": "server config log"
    },
    {
      "metric": "default_origins_logged",
      "value": "OLLAMA_ORIGINS:[http://localhost https://localhost http://localhost:* https://localhost:* http://127.0.0.1 https://127.0.0.1 http://127.0.0.1:* https://127.0.0.1:* http://0.0.0.0 https://0.0.0.0 http:",
      "unit": "text",
      "note": "server config log"
    },
    {
      "metric": "second_serve_error",
      "value": "Error: listen tcp 127.0.0.1:11434: bind: address already in use",
      "unit": "text",
      "note": "second ollama serve while the first holds :11434"
    },
    {
      "metric": "pull_ms",
      "value": 136225,
      "unit": "ms",
      "note": "ollama pull qwen2.5:0.5b wall clock on this network, n=1"
    },
    {
      "metric": "pull_tail",
      "value": "verifying sha256 digest\nwriting manifest\nsuccess",
      "unit": "text",
      "note": "last lines of ollama pull output"
    },
    {
      "metric": "repull_output",
      "value": "writing manifest\nsuccess",
      "unit": "text",
      "note": "second pull of an already-present model"
    },
    {
      "metric": "repull_ms",
      "value": 1330,
      "unit": "ms",
      "note": "re-pull of an up-to-date model (manifest check only), n=1"
    },
    {
      "metric": "list_output",
      "value": "NAME            ID              SIZE      MODIFIED\nqwen2.5:0.5b    a8b0c5157701    397 MB    Less than a second ago",
      "unit": "text",
      "note": "ollama list"
    },
    {
      "metric": "ls_alias_output",
      "value": "NAME            ID              SIZE      MODIFIED",
      "unit": "text",
      "note": "ollama ls header"
    },
    {
      "metric": "models_dir_mb",
      "value": 380,
      "unit": "MB",
      "note": "size of ~/.ollama/models after one pull"
    },
    {
      "metric": "models_dir_tree",
      "value": ". ./blobs ./manifests ./manifests/registry.ollama.ai ./manifests/registry.ollama.ai/library ./manifests/registry.ollama.ai/library/qwen2.5 ./metadata",
      "unit": "text",
      "note": "directory layout of the default model store"
    },
    {
      "metric": "blob_count",
      "value": 5,
      "unit": "count",
      "note": "blobs after one pull"
    },
    {
      "metric": "show_output",
      "value": "  Model\n    architecture        qwen2\n    parameters          494.03M\n    context length      32768\n    embedding length    896\n    quantization        Q4_K_M\n\n  Capabilities\n    completion\n    tools\n\n  System\n    You are Qwen, created by Alibaba Cloud. You are a helpful assistant.\n\n  License\n    Apache License\n    Version 2.0, January 2004\n    ...",
      "unit": "text",
      "note": "ollama show qwen2.5:0.5b"
    },
    {
      "metric": "show_parameters",
      "value": "",
      "unit": "text",
      "note": "ollama show --parameters"
    },
    {
      "metric": "show_modelfile_head",
      "value": "# Modelfile generated by \"ollama show\"\n# To build a new Modelfile based on this, replace FROM with:\n# FROM qwen2.5:0.5b\nFROM /tmp/ollama-exp/home/.ollama/models/blobs/sha256-c5396e06af294bd101b30dce59131a76d2b773e76950acc870eda801d3ab0515\nTEMPLATE \"\"\"{{- if .Messages }}\n{{- if or .System .Tools }}<|im_start|>system",
      "unit": "text",
      "note": "ollama show --modelfile (first lines)"
    },
    {
      "metric": "show_license_head",
      "value": "                                 Apache License",
      "unit": "text",
      "note": "ollama show --license"
    },
    {
      "metric": "run_oneshot_cold_ms",
      "value": 921,
      "unit": "ms",
      "note": "ollama run MODEL PROMPT including model load, n=1, Apple M1 8GB"
    },
    {
      "metric": "run_oneshot_reply",
      "value": "        ready",
      "unit": "text",
      "note": "model reply"
    },
    {
      "metric": "run_verbose_stats",
      "value": "total duration:       844.779791ms\nload duration:        779.454458ms\nprompt eval count:    34 token(s)\nprompt eval duration: 50.275ms\nprompt eval rate:     676.28 tokens/s\neval count:           2 token(s)\neval duration:        12.033ms\neval rate:            166.21 tokens/s",
      "unit": "text",
      "note": "--verbose timing block"
    },
    {
      "metric": "eval_rate_tps",
      "value": 166.21,
      "unit": "tokens/s",
      "note": "--verbose eval rate, cold run, n=1"
    },
    {
      "metric": "run_oneshot_warm_ms",
      "value": 100,
      "unit": "ms",
      "note": "same one-shot with the model already loaded, n=1"
    },
    {
      "metric": "piped_stdin_reply",
      "value": "The word \"pipe\" is a common abbreviation for \"portable air conditioner\", wh",
      "unit": "text",
      "note": "echo prompt | ollama run MODEL"
    },
    {
      "metric": "ps_loaded",
      "value": "NAME            ID              SIZE      PROCESSOR    CONTEXT    UNTIL\nqwen2.5:0.5b    a8b0c5157701    479 MB    100% GPU     4096       About a minute from now",
      "unit": "text",
      "note": "ollama ps while loaded (run used --keepalive 2m)"
    },
    {
      "metric": "api_ps",
      "value": "[{\"name\":\"qwen2.5:0.5b\",\"size\":479954205,\"size_vram\":479954205,\"context_length\":4096,\"expires_at\":\"2026-09-27T00:41:01.972195+05:00\"}]",
      "unit": "json",
      "note": "GET /api/ps"
    },
    {
      "metric": "stop_output",
      "value": "",
      "unit": "text",
      "note": "ollama stop MODEL (empty = success)"
    },
    {
      "metric": "ps_after_stop",
      "value": "NAME    ID    SIZE    PROCESSOR    CONTEXT    UNTIL",
      "unit": "text",
      "note": "ollama ps after stop"
    },
    {
      "metric": "stop_not_loaded",
      "value": "",
      "unit": "text",
      "note": "ollama stop on a model that is not loaded"
    },
    {
      "metric": "slash_help",
      "value": "Available Commands:\n  /set            Set session variables\n  /show           Show model information\n  /load <model>   Load a session or model\n  /save <model>   Save your current session\n  /clear          Clear session context\n  /bye            Exit\n  /?, /help       Help for a command\n  /? shortcuts    Help for keyboard shortcuts\n\nUse \"\"\" to begin a multi-line message.",
      "unit": "text",
      "note": "/? inside ollama run"
    },
    {
      "metric": "slash_set_parameter",
      "value": "Set parameter 'temperature' to '0.1'",
      "unit": "text",
      "note": "/set parameter temperature 0.1"
    },
    {
      "metric": "slash_set_system",
      "value": "Set system message.",
      "unit": "text",
      "note": "/set system"
    },
    {
      "metric": "slash_show_parameters",
      "value": "User defined parameters:\n  temperature                    0.1",
      "unit": "text",
      "note": "/show parameters after /set"
    },
    {
      "metric": "slash_show_info_head",
      "value": "",
      "unit": "text",
      "note": "/show info"
    },
    {
      "metric": "slash_clear",
      "value": "Cleared session context",
      "unit": "text",
      "note": "/clear"
    },
    {
      "metric": "slash_save",
      "value": "Created new model 'my-chat-snapshot'",
      "unit": "text",
      "note": "/save my-chat-snapshot"
    },
    {
      "metric": "list_after_save",
      "value": "NAME                       ID              SIZE      MODIFIED\nmy-chat-snapshot:latest    5b8ee4b69d17    397 MB    Less than a second ago\nqwen2.5:0.5b               a8b0c5157701    397 MB    10 seconds ago",
      "unit": "text",
      "note": "ollama list after /save"
    },
    {
      "metric": "set_help",
      "value": "Available Commands:\n  /set parameter ...     Set a parameter\n  /set system <string>   Set system message\n  /set history           Enable history\n  /set nohistory         Disable history\n  /set wordwrap          Enable wordwrap\n  /set nowordwrap        Disable wordwrap\n  /set format json       Enable JSON mode\n  /set noformat          Disable formatting\n  /set verbose           Show LLM stats\n  /set quiet             Disable LLM stats\n  /set think             Enable thinking\n  /set nothink           Disable thinking\n\n>>> Send a message (/? for help)/bye\n...",
      "unit": "text",
      "note": "/set with no argument lists its subcommands"
    },
    {
      "metric": "cp_output",
      "value": "copied 'qwen2.5:0.5b' to 'my-qwen'",
      "unit": "text",
      "note": "ollama cp MODEL my-qwen"
    },
    {
      "metric": "cp_disk_delta_kb",
      "value": 4,
      "unit": "KB",
      "note": "store growth from ollama cp (blobs are shared, only a manifest is written)"
    },
    {
      "metric": "list_after_cp",
      "value": "NAME              ID              SIZE      MODIFIED\nmy-qwen:latest    a8b0c5157701    397 MB    Less than a second ago\nqwen2.5:0.5b      a8b0c5157701    397 MB    15 seconds ago",
      "unit": "text",
      "note": "ollama list after cp"
    },
    {
      "metric": "modelfile_used",
      "value": "FROM qwen2.5:0.5b\nPARAMETER temperature 0.2\nPARAMETER num_ctx 8192\nSYSTEM You are a terse shell assistant. Answer with one command.",
      "unit": "text",
      "note": "Modelfile for ollama create"
    },
    {
      "metric": "create_ms",
      "value": 90,
      "unit": "ms",
      "note": "ollama create from a FROM-existing-model Modelfile, n=1"
    },
    {
      "metric": "create_output",
      "value": "creating new layer sha256:4a75c8c3b3b93eaaa643a1a6109842943c61596ff54e6b4f08356c11f30d94fa\ncreating new layer sha256:a239c746789db254b644cc1d75a26e40951b77db11195d55ba49f5223a05326b\nwriting manifest\nsuccess",
      "unit": "text",
      "note": "ollama create output"
    },
    {
      "metric": "create_show_parameters",
      "value": "num_ctx                        8192\ntemperature                    0.2",
      "unit": "text",
      "note": "ollama show shell-helper --parameters"
    },
    {
      "metric": "create_show_system",
      "value": "You are a terse shell assistant. Answer with one command.",
      "unit": "text",
      "note": "ollama show shell-helper --system"
    },
    {
      "metric": "create_run_reply",
      "value": "     To list files, including hidden ones, in a Unix-like environment, you can u\nuse the `ls` command. Here's how you can do it:",
      "unit": "text",
      "note": "reply from the custom model"
    },
    {
      "metric": "create_default_file_error",
      "value": "Error: no Modelfile or safetensors files found",
      "unit": "text",
      "note": "ollama create without -f and no ./Modelfile"
    },
    {
      "metric": "list_after_create",
      "value": "NAME                   ID              SIZE      MODIFIED\nshell-helper:latest    57248bfeaa2e    397 MB    1 second ago\nmy-qwen:latest         a8b0c5157701    397 MB    1 second ago\nqwen2.5:0.5b           a8b0c5157701    397 MB    17 seconds ago",
      "unit": "text",
      "note": "ollama list after create"
    },
    {
      "metric": "rm_output",
      "value": "deleted 'my-qwen'\ndeleted 'shell-helper'",
      "unit": "text",
      "note": "ollama rm my-qwen shell-helper"
    },
    {
      "metric": "rm_missing",
      "value": "Error: model 'does-not-exist' not found",
      "unit": "text",
      "note": "ollama rm on a missing model"
    },
    {
      "metric": "run_missing",
      "value": "Error: pull model manifest: file does not exist",
      "unit": "text",
      "note": "ollama run on a model that is not in the registry"
    },
    {
      "metric": "api_tags",
      "value": "[{\"name\":\"qwen2.5:0.5b\",\"size\":397821319,\"details\":{\"family\":\"qwen2\",\"parameter_size\":\"494.03M\",\"quantization_level\":\"Q4_K_M\"}}]",
      "unit": "json",
      "note": "GET /api/tags"
    },
    {
      "metric": "api_generate_keys",
      "value": "[\"context\",\"created_at\",\"done\",\"done_reason\",\"eval_count\",\"eval_duration\",\"load_duration\",\"model\",\"prompt_eval_cached_count\",\"prompt_eval_count\",\"prompt_eval_duration\",\"response\",\"total_duration\"]",
      "unit": "json",
      "note": "top-level keys of a non-streaming /api/generate response"
    },
    {
      "metric": "api_generate_eval_count",
      "value": 36,
      "unit": "tokens",
      "note": "eval_count from /api/generate"
    },
    {
      "metric": "api_generate_total_ms",
      "value": 937,
      "unit": "ms",
      "note": "total_duration from /api/generate (ns converted to ms), n=1"
    },
    {
      "metric": "api_generate_response",
      "value": "The sky appears blue because it is primarily composed of water vapor, which absorbs and scatters the blue light from the sun, creating the visible colors we see in the atmosphere.",
      "unit": "text",
      "note": "response text"
    },
    {
      "metric": "api_ps_after_keepalive0",
      "value": 0,
      "unit": "count",
      "note": "models loaded after a request with keep_alive:0"
    },
    {
      "metric": "api_show_ctx",
      "value": 32768,
      "unit": "tokens",
      "note": "model_info context_length from /api/show"
    },
    {
      "metric": "openai_models",
      "value": "[\"qwen2.5:0.5b\"]",
      "unit": "json",
      "note": "GET /v1/models (OpenAI-compatible)"
    },
    {
      "metric": "openai_chat_raw",
      "value": "{\"id\":\"chatcmpl-904\",\"object\":\"chat.completion\",\"created\":1790451560,\"model\":\"qwen2.5:0.5b\",\"system_fingerprint\":\"fp_ollama\",\"choices\":[{\"index\":0,\"message\":{\"role\":\"assistant\",\"content\":\"Ok! How can I assist you today?\"},\"finish_reason\":\"stop\"}],\"usage\":{\"prompt_tokens\":31,\"prompt_tokens_details\":{\"cached_tokens\":0},\"completion_tokens\":10,\"total_tokens\":41}}",
      "unit": "json",
      "note": "POST /v1/chat/completions raw body"
    },
    {
      "metric": "openai_chat_object",
      "value": "chat.completion",
      "unit": "text",
      "note": "object field of /v1/chat/completions"
    },
    {
      "metric": "lan_request_result",
      "value": "code_000",
      "unit": "text",
      "note": "request to this machine's LAN IP while bound to 127.0.0.1 (000 = refused)"
    },
    {
      "metric": "alt_listen_line",
      "value": "Listening on 127.0.0.1:11500",
      "unit": "text",
      "note": "serve log with OLLAMA_HOST=127.0.0.1:11500"
    },
    {
      "metric": "alt_list_output",
      "value": "NAME    ID    SIZE    MODIFIED",
      "unit": "text",
      "note": "OLLAMA_HOST=127.0.0.1:11500 ollama list (separate OLLAMA_MODELS dir)"
    },
    {
      "metric": "default_list_still",
      "value": 2,
      "unit": "lines",
      "note": "ollama list against :11434 at the same time (header + models)"
    },
    {
      "metric": "alt_models_dir_created",
      "value": ". ./blobs ./manifests",
      "unit": "text",
      "note": "contents of the custom OLLAMA_MODELS dir on startup"
    },
    {
      "metric": "client_scheme_form",
      "value": "ollama version is 0.34.4",
      "unit": "text",
      "note": "client with OLLAMA_HOST including http:// scheme"
    },
    {
      "metric": "rm_base_output",
      "value": "deleted 'qwen2.5:0.5b'",
      "unit": "text",
      "note": "ollama rm qwen2.5:0.5b"
    },
    {
      "metric": "models_dir_after_rm_kb",
      "value": 0,
      "unit": "KB",
      "note": "store size after removing the only model"
    },
    {
      "metric": "list_after_rm",
      "value": "NAME    ID    SIZE    MODIFIED",
      "unit": "text",
      "note": "ollama list after rm"
    },
    {
      "metric": "home_dot_ollama",
      "value": ". ./cache ./history ./id_ed25519 ./id_ed25519.pub ./models",
      "unit": "text",
      "note": "files Ollama created in ~/.ollama"
    }
  ],
  "log": [
    "[harness] starting ollama serve on default 127.0.0.1:11434",
    "[harness] pulling qwen2.5:0.5b",
    "[harness] one-shot run",
    "[harness] interactive session",
    "spawn /tmp/ollama-exp/bin/ollama run qwen2.5:0.5b",
    "\u001b]11;?\u001b\\>>> Send a message (/? for help)/?",
    "Available Commands:",
    "  /set            Set session variables",
    "  /show           Show model information",
    "  /load <model>   Load a session or model",
    "  /save <model>   Save your current session",
    "  /clear          Clear session context",
    "  /bye            Exit",
    "  /?, /help       Help for a command",
    "  /? shortcuts    Help for keyboard shortcuts",
    "",
    "Use \"\"\" to begin a multi-line message.",
    "",
    ">>> Send a message (/? for help)/set",
    "...  par",
    "... amet",
    "... er t",
    "... empe",
    "... ratu",
    "... re 0",
    "... .1",
    "Set parameter 'temperature' to '0.1'",
    ">>> Send a message (/? for help)/set",
    "...  sys",
    "... tem ",
    "... You ",
    "... answ",
    "... er i",
    "... n fi",
    "... ve w",
    "... ords",
    "...  or ",
    "... fewe",
    "... r.",
    "Set system message.",
    ">>> Send a message (/? for help)/sho",
    "... w pa",
    "... rame",
    "... ters",
    "... ",
    "Model defined parameters:",
    "  No additional parameters were specified for this model.",
    "",
    "User defined parameters:",
    "  temperature                    0.1",
    "",
    ">>> Send a message (/? for help)/sho",
    "... w in",
    "... fo",
    "  Model",
    "    architecture        qwen2      ",
    "    parameters          494.03M    ",
    "    context length      32768      ",
    "    embedding length    896        ",
    "    quantization        Q4_K_M     ",
    "",
    "  Capabilities",
    "    completion    ",
    "    tools         ",
    "",
    "  System",
    "    You answer in five words or fewer.    ",
    "",
    "  License",
    "    Apache License               ",
    "    Version 2.0, January 2004    ",
    "    ...                          ",
    "",
    ">>> Send a message (/? for help)/set",
    "...  ver",
    "... bose",
    "... ",
    "Set 'verbose' mode.",
    ">>> Send a message (/? for help)/cle",
    "... ar",
    "Cleared session context",
    ">>> Send a message (/? for help)/loa",
    "... d qw",
    "... en2.",
    "... 5:0.",
    "... 5b",
    "Loading model 'qwen2.5:0.5b'",
    ">>> Send a message (/? for help)/sav",
    "... e my",
    "... -cha",
    "... t-sn",
    "... apsh",
    "... ot",
    "Created new model 'my-chat-snapshot'",
    ">>> Send a message (/? for help)/bye",
    "... ",
    "[harness] second server on 11500 with OLLAMA_MODELS=/tmp/ollama-exp/models-alt",
    "[harness] done",
    "[harness] cleanup",
    ""
  ]
}