harvardMadsys/freeinference_agentic_trace

FreeInference Agentic Trace

This dataset contains 16 weeks (2026-05-17 to 2026-09-05) of coding and assistant agents interacting with LLMs through the FreeInference gateway: 12,002 agent sessions from 267 accounts and 14 agent harnesses, with 1,186,582 LLM requests and 1,213,347 tool calls.

The dataset accompanies the paper From Requests to Sessions: A Large-Scale Characterization of Human-Driven Agentic Workloads. The code that reads it, replays its prefix cache and reproduces the paper's figures is at HarvardMadSys/freeinference_agentic_trace.

The trace includes request timing and serving latency, models and anonymized providers, token usage, tool definitions and calls, input mutations, and block-level prefixes for cache replay. No message text is released.

Traces

The primary release is under:

release/<week>/traces/<day>.jsonl.zst

Each file is zstd-compressed JSONL with one agent session per line. A session contains its LLM requests in order. Each request records the model invocation and the tool calls produced by the response. If a tool call launches a subagent, the complete subagent session is nested under that call.

A trace has the following structure:

{
  "week": "2026-05-17_2026-05-23",
  "day": "2026-05-17",
  "format_version": 1,

  // One reconstructed agent session.
  "session": {
    "session_id": "...",
    "user_id": "...",
    "harness": "claude-code",
    "start_ms": 1778977459236,
    "end_ms": 1778984067198,

    // LLM requests in chronological order.
    "requests": [
      {
        // Position in the session and the user turn it belongs to.
        "step": 0,
        "turn_index": 0,

        // What triggered this request and how it was linked
        // to the previous request.
        "step_trigger": "user",
        "link_reason": null,

        // Gateway timing.
        "start_ms": 1778977459236,
        "end_ms": 1778977468828,
        "ttft_ms": 7889,

        // Requested model and anonymized upstream provider.
        "model": "glm-5.1",
        "provider": "B",
        "stream": true,

        // Token counts reported by the provider.
        "provider_tokens": {
          "prompt": 52731,
          "completion": 230,
          "cached": 1280
        },

        // Input tokens broken down by message role.
        // Tokenized by tiktoken `o200k_base`
        "role_tokens": {
          "system": 1790,
          "user": 1110,
          "assistant": 15916,
          "tool": 25725,
          "tool_definitions": 8638
        },

        // Tool definitions by the harness
        "n_messages": 179,
        "tool_definition_names": [
          "read_file",
          "write_file",
          "terminal",
          "execute_code"
        ],

        "finish_reason": "tool_calls",

        // How the input changed from the previous request.
        "mutation": null,

        // Input represented as chained 16-token prefix blocks.
        // Tokenized by tiktoken `o200k_base`. IDs are local to each week.
        "block_ids": [123, 456, 789, "..."],

        // Tool calls produced by this response.
        "tool_calls": [
          {
            "call_index": 0,
            "name": "terminal",

            // Normalized tool category. Shell calls also include
            // the programs identified in the command.
            "category": "*File Ops",
            "programs": ["ps", "echo"],
            "parse_status": "success",

            // Arguments are anonymized while preserving their structure.
            "arguments": [
              [
                "command",
                "ps -p <num> -o <str> 2> <path> || echo <str>"
              ]
            ],

            // Tool result and time until the next LLM request.
            "outcome": "ok",
            "result_tokens": 27,
            "latency_ms": 8025,

            // A spawned subagent appears as another complete session.
            "subagent": null
          }
        ]
      },

      {
        "step": 1,
        "turn_index": 0,
        "step_trigger": "tool",
        "link_reason": "tool_call_id",

        // Subsequent requests contain the same fields.
        // mutation records whether the input was appended to,
        // modified, or switched relative to the previous request.
        "mutation": {
          "transition": "append",
          "cause": null
        },

        "tool_calls": ["..."]
      }
    ]
  }
}

For convenience, processed/ contains Parquet tables derived by parsing and flattening the traces into sessions, requests, and tool calls.

Notes

• Request inputs are tokenized with o200k_base and split into 16-token blocks. A partial block is dropped.
• Sessions and block IDs are local to each week. When replaying multiple weeks together, make block IDs unique across weeks, for example by prefixing each ID with the week.
• Sessions do not cross week boundaries. A session that continues into the next week is released as a new session.
• Sessions are reconstructed by linking related requests using repeated tool-call IDs and assistant messages.
• Rare tool names and commands used by fewer than 3 accounts are replaced with other or a typed placeholder.
• provider_tokens are reported by the upstream provider, while other token counts are recomputed using o200k_base, so their values may differ.

Usage

DuckDB

Query the derived Parquet tables directly from the Hugging Face Hub:

SELECT
    sessions.harness,
    count(*) AS requests,
    approx_quantile(requests.ttft_ms, 0.5) AS p50_ttft_ms
FROM 'hf://datasets/harvardMadsys/freeinference_agentic_trace/release/*/processed/requests.parquet' AS requests
JOIN 'hf://datasets/harvardMadsys/freeinference_agentic_trace/release/*/processed/sessions.parquet' AS sessions
    USING (session_id)
GROUP BY sessions.harness
ORDER BY requests DESC;

Hugging Face Datasets

from datasets import load_dataset

traces = load_dataset(
    "harvardMadsys/freeinference_agentic_trace",
    "traces",
    split="week_2026_08_30",
)

The derived request and tool-call tables are also available as separate configurations for convenience:

requests = load_dataset(
    "harvardMadsys/freeinference_agentic_trace",
    "requests",
    split="week_2026_08_30",
    streaming=True,
)

tool_calls = load_dataset(
    "harvardMadsys/freeinference_agentic_trace",
    "tool_calls",
    split="week_2026_08_30",
    streaming=True,
)

Download

Download the full dataset:

hf download harvardMadsys/freeinference_agentic_trace \
    --repo-type dataset \
    --local-dir freeinference_trace

Or download one week:

hf download harvardMadsys/freeinference_agentic_trace \
    --repo-type dataset \
    --local-dir freeinference_trace \
    --include "release/2026-08-30_2026-09-05/*"

License

This dataset is released under the Creative Commons Attribution 4.0 International License (CC BY 4.0).

If you use this dataset, please cite the paper below.

Citation

If you use the dataset or artifact in your research, please cite our paper:

@article{freeinference2026requests,
  title={From Requests to Sessions: A Large-Scale Characterization of Human-Driven Agentic Workloads},
  author={Nixon, William and Tian, Muxin and Zheng, Yunjia and Gunawi, Haryadi S. and Yang, Juncheng},
  year={2026}
}