diff --git a/CHANGELOG.md b/CHANGELOG.md
index 965a8fac..e5544bea 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -4,6 +4,17 @@ User-visible changes to Imp are recorded here.
 
 ## Unreleased
 
+## 0.4.0 — 2026-09-17
+
+- `Imp.ACP` and `Imp.MCP.connect/2` are part of Imp. The separate `imp_acp`
+  package is retired: `Imp.ACP.start_link/1` and `Imp.ACP.run/1` expose an
+  ordinary Imp program to an ACP host, and `Imp.MCP.connect/2` imports
+  authorized MCP servers through ExMCP as ordinary tools with explicit
+  connection cleanup. There is no compatibility shim; a consumer that depended
+  on `imp_acp` depends on `imp` alone and changes the module prefix. ExMCP is
+  declared `runtime: false`, so an OTP release that uses either must list
+  `applications: [ex_mcp: :load]` in its release definition. Ordinary Imp
+  startup still starts no protocol endpoint.
 - A map or list value in a prompt, including a structured tool result, now
   renders the way DSPy renders a dict: `json.dumps(..., ensure_ascii=False)`
   with Python's default separators, complete. It was `inspect/1` at its
@@ -88,6 +99,24 @@ User-visible changes to Imp are recorded here.
   `{:mcp_tools_list_failed, server, {:invalid_mcp_tools_response, shape}}`
   rather than as a bare `{:invalid_mcp_tools_response, shape}` that named no
   server.
+- The repository is public. Installing at a tag needs no credentials, and the
+  README, docs, and livebooks no longer describe a private source release.
+- Removed the evidence-certification bookkeeping from the source checkout. It
+  never shipped in the package, so a consumer sees no change; the benchmark
+  harness it wrapped is unchanged.
+- Added [Benchmarks](https://github.com/deepfates/imp/blob/main/docs/BENCHMARKS.md)
+  and its [results table](https://github.com/deepfates/imp/blob/main/benchmarks/RESULTS.md).
+  Every number this repository publishes is one row in that table, carrying
+  the dataset and its license, the model, the provider, the date, the commit,
+  and the command that produced it; prose elsewhere cites a row rather than
+  restating a number. The benchmarks page says what each command needs from
+  you — key, Python environment, time, rough cost — and separates a row a
+  stranger can re-measure with an API key from one that only recomputes
+  statistics from committed rows, and from the claims that cannot be
+  re-measured at all. Neither page publishes an aggregate or a release score.
+  The tutorial's rows were re-measured live for this release, a month after the
+  first run, and both runs are recorded. They live in the repository, not in
+  the installed package.
 
 ## 0.3.2 — 2026-09-01
 
diff --git a/README.md b/README.md
index 56054a58..b725dbc0 100644
--- a/README.md
+++ b/README.md
@@ -6,12 +6,6 @@ ordinary OTP application. It brings the central idea of
 [DSPy](https://dspy.ai)—programming behavior and improving it from examples
 rather than hand-editing prompts—to the BEAM.
 
-**Release state.** The released tag `v0.3.2` does not contain `Imp.ACP`. The
-integrated protocol adapters — `Imp.ACP` and `Imp.MCP.connect/2`, absorbed from
-the retired `imp_acp` package — exist only on `main` until the next tag.
-Installing `v0.3.2` gives you the library described below without them; using
-the adapters today means depending on `main` (see [Install](#install)).
-
 Here, “typed” means required inputs are checked and model outputs are parsed
 and validated against the signature before application code receives them.
 For DSPy compatibility, a supplied input whose value disagrees with its
@@ -135,13 +129,9 @@ Imp is not published to Hex. It is installed from this public repository, at
 an immutable tag, with no credentials:
 
 ```elixir
-{:imp, github: "deepfates/imp", tag: "v0.3.2"}
+{:imp, github: "deepfates/imp", tag: "v0.4.0"}
 ```
 
-That tag does not contain `Imp.ACP` or `Imp.MCP.connect/2`; a tag that does
-will be cut. Until then, depend on `{:imp, github: "deepfates/imp", branch:
-"main"}` if you need the adapters.
-
 ExMCP is declared `runtime: false`, so an OTP release that uses `Imp.ACP` or
 `Imp.MCP` must list `applications: [ex_mcp: :load]` in its release definition;
 see [protocol runtime in releases](docs/PRODUCTION_OPERATIONS.md#protocol-runtime-in-releases).
@@ -152,9 +142,9 @@ modules are ported from.
 Use `{:imp, path: "path/to/imp"}` only while developing against a local
 checkout. Imp requires Elixir `~> 1.19`. Commit your application's `mix.lock`;
 the Git tag fixes Imp's source, while normal Mix constraints may otherwise
-resolve newer compatible transitive versions. Version `0.3.2` contains the
-breaking `0.3` changes from `0.2.1`; see the [release notes](RELEASE_NOTES.md)
-when upgrading.
+resolve newer compatible transitive versions. Version `0.4.0` contains
+breaking changes from `0.3.2`; see the [release notes](RELEASE_NOTES.md) when
+upgrading.
 
 Imp uses [ReqLLM](https://hex.pm/packages/req_llm) for model providers. The
 examples use OpenAI, but programs are not tied to that provider. The
@@ -164,9 +154,6 @@ provider-free parts of the learning path and deployment example do too.
 
 ## Connect tools or expose a program
 
-Everything in this section is on `main` only, not in `v0.3.2` (see the release
-state at the top).
-
 `Imp.MCP.connect/2` imports authorized MCP servers through ExMCP, returning
 ordinary tools plus explicit connection cleanup. Source server/tool identities,
 schemas, and annotations remain in each tool's `metadata.mcp` even when names
@@ -209,8 +196,8 @@ Run `mix docs` for the exhaustive module and function reference.
 
 For an ordinary ACP workspace agent with bounded tools, see
 [examples/workspace_agent](examples/workspace_agent/README.md). It depends
-directly on this Imp checkout (`main` only, like the adapters it uses) and
-includes a provider-free mode for checking its launcher and workspace boundary.
+directly on this Imp checkout by path and includes a provider-free mode for
+checking its launcher and workspace boundary.
 
 ## Where this fits
 
diff --git a/RELEASE_NOTES.md b/RELEASE_NOTES.md
index fc3be01f..02caffdc 100644
--- a/RELEASE_NOTES.md
+++ b/RELEASE_NOTES.md
@@ -1,101 +1,78 @@
-# Imp v0.3.2
+# Imp v0.4.0
 
 Imp is a framework for typed, optimizable language-model programs on the BEAM.
 Declare a task as named inputs and outputs, call it like any other Elixir
 program, measure it on examples, compile it with an optimizer, and run the
 selected program under OTP.
 
+This release absorbs the protocol adapters that previously lived on `main`
+only, and changes three published shapes. It is `0.4.0` rather than a patch
+because a program written against `v0.3.2` can need edits.
+
 ## Install
 
-`v0.3.2` is a Git source release from a public repository; no credentials are
+`v0.4.0` is a Git source release from a public repository; no credentials are
 required.
 
 ```elixir
-{:imp, github: "deepfates/imp", tag: "v0.3.2"}
+{:imp, github: "deepfates/imp", tag: "v0.4.0"}
 ```
 
 Imp is not published to Hex. Use a path dependency only while developing
 against a local checkout.
 
-This patch release makes the RLM controller's language guide match its
-restricted interpreter and preserves explicit zero-retry provider policy for
-all Req adapter forms, including MFA adapters with arguments.
-
-## What is included
-
-- Typed signatures with scalar, collection, enum, union, optional, default,
-  code, and constrained fields.
-- `Predict`, `ChainOfThought`, composed `Imp.Module` programs, retrieval,
-  ReActV2, CodeAct, RLM, tools, MCP, and provider streaming.
-- Examples, metrics, concurrent evaluation, disjoint train/selection/test
-  experiments, and optimizer reports.
-- Demonstration, instruction, prompt, ensemble, rule, playbook, and
-  weight-training optimizer families, including GEPA, MIPROv2, SIMBA, COPRO,
-  BootstrapFewShot, RandomSearch, KNNFewShot, BootstrapFinetune, BetterTogether,
-  Avatar, and Optimize Anything.
-- Checksummed whole-program and parameter artifacts that exclude credentials
-  and apply selected state to freshly constructed trusted code.
-- OTP-native operation with bounded tasks, cancellation, per-effect
-  authorization, redacted telemetry, caching, usage accounting, hot reload,
-  and failure propagation.
-- ReqLLM provider clients, explicit local/static test models, retriever and
-  trainer extension points, and local MLX/TRL integration boundaries.
-
-The [Learning Path](docs/LEARNING_PATH.md) builds one program from its first
-provider call through evaluation, optimization, tools, persistence, and
-deployment. The [deployment example](examples/deployment/README.md) shows a
-supervised two-stage program with parameter reload, concurrent calls, restart,
-timeouts, and crash containment. Five Livebooks cover the same system
-interactively.
-
-## What the BEAM changes
-
-Imp preserves DSPy's program/evaluate/optimize workflow without copying
-Python's object model. Programs are immutable values. Configuration can be
-explicit or process-scoped. Evaluation and tool work run in supervised tasks.
-Telemetry uses standard `:telemetry` events. Saved state is rebound to live
-providers and callbacks at application startup instead of serializing runtime
-authority.
-
-Provider output and optimizer search are stochastic. A compiled program is a
-candidate until it improves the metric that matters on data excluded from
-training and selection. Imp supplies that lifecycle; applications still own
-their data, metric, budget, promotion rule, and operational policy.
-
-DSPy's Python integration ecosystem is larger. Imp exposes extension points
-for providers, retrievers, adapters, tools, and trainers, but Python-only
-integrations do not automatically work on the BEAM. DSPy's Flex code optimizer
-is not included in this release.
-
-The supported center is the `Imp` facade, signatures, adapters, evaluation,
-static and ReqLLM execution, tools, telemetry, saving, and the deployment
-pattern. Generated docs place optimizer implementations, parameter artifacts,
-agent loops, training integrations, and `Imp.Run` in **Experimental optimizers
-and advanced workflows**. These are implemented and tested APIs, not release
-promises of effectiveness or pre-1.0 shape stability. In particular, GRPO is
-an external-training boundary rather than an in-process gradient engine.
-
-## Breaking changes from v0.2.1
-
-- `Imp.optimize/3`, `/4`, and `/5` return `{:ok, program}` or
-  `{:error, reason}`. Use the corresponding `Imp.optimize!` function when a
-  failure should raise.
-- `Imp.Adapters.Types` and its nested structs moved to `Imp.Adapter.Types`.
-- `Imp.Agent` and `Imp.Agent.Runtime` were removed. Use ReActV2 or RLM as the
-  program and ordinary Elixir supervision as the runtime. Use
-  `Imp.start_run/3` only when a host needs ordered events, addressable
-  cancellation, or explicit effect authorization.
+ExMCP is declared `runtime: false`, so an OTP release that uses `Imp.ACP` or
+`Imp.MCP` must list `applications: [ex_mcp: :load]` in its release
+definition; see [protocol runtime in
+releases](docs/PRODUCTION_OPERATIONS.md#protocol-runtime-in-releases).
+Ordinary Imp startup starts no protocol endpoint.
+
+## Headline changes
+
+- `Imp.ACP` and `Imp.MCP.connect/2` are in the tag. The separate `imp_acp`
+  package is retired with no compatibility shim: a consumer that depended on
+  it now depends on `imp` alone. `Imp.MCP.connect/2` also gains OAuth
+  credentials for remote HTTP servers (`Imp.MCP.OAuth`), `bearer_env`
+  descriptor auth, `on_failure: :drop` with an `unavailable` list, and a
+  per-dial timeout.
+- `:reasoning_effort` is the one reasoning option on `Imp.Clients.ReqLLM`.
+  `:openrouter_reasoning` is gone; the wire encoding is the separate
+  `:openrouter_reasoning_wire`.
+- ReActV2 sends the tool roster natively and no longer declares a `tools`
+  input field or writes its instructions into `signature.instructions`. Loop
+  guidance travels to the adapter through `:adapter_opts`.
+- The `:model_response` event's `metadata.cost` is a plain USD float or `nil`,
+  with any provider breakdown under `metadata.billing`.
+- Structured values in a prompt render complete, the way DSPy renders a dict,
+  instead of a truncated `inspect/1`.
+
+## Breaking changes from v0.3.2
+
+- Replace `openrouter_reasoning: ...` with `reasoning_effort: ...`. Saved
+  programs allowlist `:reasoning_effort` and `:openrouter_reasoning_wire` in
+  its place, so rebuild artifacts that carried the old key.
+- A caller that passed or read ReActV2's `tools` input field no longer has
+  one; the roster is sent natively.
+- A host that read `metadata.cost` as a provider billing map reads a number
+  now, and finds the map under `metadata.billing` when the provider sent one.
 
 ## Upgrade path
 
-1. Replace `Imp.Adapters.Types` references with `Imp.Adapter.Types`.
-2. Choose the returning or raising optimizer API explicitly.
-3. Replace `Imp.Agent` usage with a ReActV2/RLM program owned by your
-   supervision tree.
-4. Rebuild saved artifacts with `0.3.2` before promotion.
+1. Rename the reasoning option and rebuild saved artifacts with `0.4.0`.
+2. Drop any `tools` handling around ReActV2.
+3. Sum spend from `metadata.cost` as a number.
+4. If you depended on `{:imp, github: "deepfates/imp", branch: "main"}` for the
+   adapters, move to the tag.
 5. Run your held-out evaluation and application smoke test against the tagged
    dependency.
 
-Generated module documentation is the complete API reference. Start with
-`Imp`, `Imp.Signature`, `Imp.Module`, `Imp.Evaluate`, `Imp.Optimizer`,
-`Imp.Optimizer.Artifact`, `Imp.Run`, and `Imp.Telemetry`.
+New in this release: [Benchmarks](https://github.com/deepfates/imp/blob/main/docs/BENCHMARKS.md)
+and its [results table](https://github.com/deepfates/imp/blob/main/benchmarks/RESULTS.md)
+carry every number this repository publishes with the command that produces it,
+and the ticket-routing rows were re-measured live for this release, a month
+after the first run, with both runs recorded.
+
+The [CHANGELOG](CHANGELOG.md) records every user-visible change in this
+release. Generated module documentation is the complete API reference. Start
+with `Imp`, `Imp.Signature`, `Imp.Module`, `Imp.Evaluate`, `Imp.Optimizer`,
+`Imp.ACP`, `Imp.MCP`, and `Imp.Telemetry`.
diff --git a/benchmarks/RESULTS.md b/benchmarks/RESULTS.md
index 02fea676..8d0b130b 100644
--- a/benchmarks/RESULTS.md
+++ b/benchmarks/RESULTS.md
@@ -18,19 +18,36 @@ for the claims that cannot be re-measured at all.
 
 | # | Number | Dataset (license) | Model | Provider | Date | Commit | Command |
 | --- | --- | --- | --- | --- | --- | --- | --- |
-| R1 | Zero-shot held-out accuracy `0.30`–`0.50` over 3 repeats (20 held-out of 60 tickets) | `priv/tutorial/support_tickets.json`, 60 rows, sha256 `7ea5ae7a…` (written for this repository; see [SUPPORT_TICKETS_LICENSE.md](../priv/tutorial/SUPPORT_TICKETS_LICENSE.md)) | `gpt-5.4-mini` | OpenRouter route `openai/gpt-5.4-mini` | 2026-08-22 | `88d61a9c` | `OPENAI_API_KEY=… mix run scripts/tutorial_ticket_routing_experiment.exs` |
-| R2 | `LabeledFewShot(k: 8)` held-out accuracy `0.95`–`1.00` over the same 3 repeats; per-repeat lift `+0.45`, `+0.65`, `+0.65` | same as R1 | `gpt-5.4-mini` | OpenRouter route `openai/gpt-5.4-mini` | 2026-08-22 | `88d61a9c` | same as R1 |
+| R1 | Zero-shot held-out accuracy `0.30`–`0.40` over 3 repeats (20 held-out of 60 tickets) | `priv/tutorial/support_tickets.json`, 60 rows, sha256 `7ea5ae7a…` (written for this repository; see [SUPPORT_TICKETS_LICENSE.md](../priv/tutorial/SUPPORT_TICKETS_LICENSE.md)) | `gpt-5.4-mini` | OpenRouter route `openai/gpt-5.4-mini` | 2026-09-17 | `7985ed2f` | `OPENAI_API_KEY=… mix run scripts/tutorial_ticket_routing_experiment.exs` |
+| R2 | `LabeledFewShot(k: 8)` held-out accuracy `0.90`–`0.95` over the same 3 repeats; per-repeat lift `+0.55`, `+0.55`, `+0.65` | same as R1 | `gpt-5.4-mini` | OpenRouter route `openai/gpt-5.4-mini` | 2026-09-17 | `7985ed2f` | same as R1 |
 
 R1 and R2 come from one execution of one command; they are two numbers from the
 same three repeats, not independent measurements. That run used 120 requests,
-44,365 tokens and `$0.038819` in provider-priced usage for all three repeats —
-about `$0.013` and 8–9 seconds per repeat. No row errored and the in-BEAM cache
-was cleared before each repeat, so all 120 calls were live.
+44,293 tokens and `$0.038488` in provider-priced usage for all three repeats —
+about `$0.013` and 8–13 seconds per repeat. No row errored and the in-BEAM cache
+was cleared before each repeat, so all 120 calls were live. The run's artifact
+is committed at
+[`benchmarks/data/tutorial-ticket-routing-2026-09-17.receipt.json`](data/tutorial-ticket-routing-2026-09-17.receipt.json),
+sha256 `2dcc1228…`; the dollar figure in it is the script's own pricing table
+(`$0.75`/`$4.50` per million tokens) applied to reported token counts, not a
+provider-billed amount, so treat it as an estimate and the token counts as the
+measurement.
+
+The same command was run a month earlier, on 2026-08-22 at commit `88d61a9c`,
+against the same dataset, model and route: zero-shot `0.30`–`0.50`, optimized
+`0.95`–`1.00`, per-repeat lift `+0.45`, `+0.65`, `+0.65`, 120 requests and
+`$0.038819` in provider-priced usage. Two independent runs a month apart agree
+on the thing worth claiming — every repeat improved, by 45 to 65 points — and
+disagree on the endpoints, which is what a twenty-row evaluation should do.
 
 Three repeats of a twenty-row evaluation is a coarse instrument. The gap between
-the two rows (45–65 points) is far larger than the instrument's resolution
-(one row is 5 points), which is why the direction is trustworthy while the exact
-endpoints are not.
+the two rows (55–65 points on 2026-09-17, 45–65 points on 2026-08-22) is far
+larger than the instrument's resolution (one row is 5 points), which is why the
+direction is trustworthy while the exact endpoints are not. The 2026-09-17 run
+put one optimized repeat at `0.90`, below the `0.95`–`1.00` the tutorial claimed
+from the first run alone, and one repeat at 13.1 seconds, above the 8–9 seconds
+it claimed; both claims were widened to the measured union rather than restated
+from the luckier run.
 
 ## Recomputable only
 
diff --git a/benchmarks/data/tutorial-ticket-routing-2026-09-17.receipt.json b/benchmarks/data/tutorial-ticket-routing-2026-09-17.receipt.json
new file mode 100644
index 00000000..070aaab4
--- /dev/null
+++ b/benchmarks/data/tutorial-ticket-routing-2026-09-17.receipt.json
@@ -0,0 +1,209 @@
+{
+  "configuration": {
+    "adapter": "Imp.Adapter.JSON",
+    "cache_cleared_per_repeat": true,
+    "json_retries": 1,
+    "k": 8,
+    "max_concurrency": 8,
+    "metric": "Imp.exact_match(:team)",
+    "optimizer": "Imp.Optimizer.LabeledFewShot",
+    "repeats": 3,
+    "signature": "ticket -> team: enum[atlas,harbor,beacon,quill]",
+    "timeout_ms": 60000,
+    "trainset": "train split only; dev split unused; test split held out"
+  },
+  "dataset": {
+    "dev": 20,
+    "path": "priv/tutorial/support_tickets.json",
+    "sha256": "7ea5ae7ad724dc4311cf954bafeb48a4b7b284b619c33bc00b30d37ad6df6f87",
+    "test": 20,
+    "train": 20
+  },
+  "doc_claims_under_test": {
+    "baseline_repeat_range": [
+      0.3,
+      0.5
+    ],
+    "cost_usd_per_repeat_approx": 0.013,
+    "duration_seconds_range": [
+      8,
+      9
+    ],
+    "optimized_repeat_range": [
+      0.95,
+      1.0
+    ],
+    "source": "docs/TUTORIAL_TICKET_ROUTING.md"
+  },
+  "fresh_service": {
+    "artifact_sha256": "cacd6fb5f6538da5fb7a3e7569e052e4735cafb8b404ab4e6aea3dda92537434",
+    "budget": {
+      "active_reservations": 0,
+      "exhausted": null,
+      "limits": {
+        "input_tokens": 50000,
+        "output_tokens": 1024,
+        "requests": 4,
+        "usd": 0.1
+      },
+      "pricing": {
+        "input_per_million": 0.75,
+        "output_per_million": 4.5,
+        "source_url": "https://openai.com/api/pricing/"
+      },
+      "requests": 4,
+      "reservation_ledger_version": 2,
+      "reservations": [],
+      "reserved": {
+        "input_tokens": 0,
+        "output_tokens": 0,
+        "usd": 0.0
+      },
+      "single_attempt_transport_enforced": true,
+      "transport_attempts": 4,
+      "usage": {
+        "input_tokens": 1987,
+        "output_tokens": 55,
+        "usd": 0.0017360000000000001
+      }
+    },
+    "concurrency": 4,
+    "correct": 4,
+    "fresh_os_process": true,
+    "total": 4
+  },
+  "generated_at": "2026-09-17T20:14:18Z",
+  "git_sha": "7985ed2fed47d742c45639776b8da5e6f256cc38",
+  "model": "openai:gpt-5.4-mini",
+  "provider_route": "openrouter:openai/gpt-5.4-mini",
+  "runner": "tutorial-ticket-routing-experiment",
+  "runs": [
+    {
+      "absolute_lift": 0.55,
+      "baseline_errors": 0,
+      "baseline_score": 0.35,
+      "cache": {
+        "bypasses": 0,
+        "cleared_before_run": true,
+        "hits": 0,
+        "misses": 0
+      },
+      "duration_ms": 9287,
+      "heldout_examples": 20,
+      "optimized_errors": 0,
+      "optimized_score": 0.9,
+      "run": 1,
+      "started_at": "2026-09-17T20:13:44Z",
+      "usage": {
+        "input_tokens": 14296,
+        "output_tokens": 466,
+        "requests_with_usage": 40,
+        "total_tokens": 14762
+      }
+    },
+    {
+      "absolute_lift": 0.55,
+      "baseline_errors": 0,
+      "baseline_score": 0.4,
+      "cache": {
+        "bypasses": 0,
+        "cleared_before_run": true,
+        "hits": 0,
+        "misses": 0
+      },
+      "duration_ms": 13135,
+      "heldout_examples": 20,
+      "optimized_errors": 0,
+      "optimized_score": 0.95,
+      "run": 2,
+      "started_at": "2026-09-17T20:13:53Z",
+      "usage": {
+        "input_tokens": 14296,
+        "output_tokens": 453,
+        "requests_with_usage": 40,
+        "total_tokens": 14749
+      }
+    },
+    {
+      "absolute_lift": 0.65,
+      "baseline_errors": 0,
+      "baseline_score": 0.3,
+      "cache": {
+        "bypasses": 0,
+        "cleared_before_run": true,
+        "hits": 0,
+        "misses": 0
+      },
+      "duration_ms": 8036,
+      "heldout_examples": 20,
+      "optimized_errors": 0,
+      "optimized_score": 0.95,
+      "run": 3,
+      "started_at": "2026-09-17T20:14:06Z",
+      "usage": {
+        "input_tokens": 14296,
+        "output_tokens": 486,
+        "requests_with_usage": 40,
+        "total_tokens": 14782
+      }
+    }
+  ],
+  "schema_version": 2,
+  "scope": {
+    "claimed": "LabeledFewShot(k: 8) held-out lift on the shipped sixty-ticket routing task with openrouter:openai/gpt-5.4-mini across 3 live repeats",
+    "not_claimed": [
+      "generalization beyond the shipped support-ticket dataset",
+      "search-optimizer (RandomSearch/MIPROv2) effectiveness",
+      "DSPy-matched comparison"
+    ]
+  },
+  "script": "scripts/tutorial_ticket_routing_experiment.exs",
+  "summary": {
+    "all_runs_improved": true,
+    "baseline_score_max": 0.4,
+    "baseline_score_min": 0.3,
+    "duration_ms_max": 13135,
+    "duration_ms_min": 8036,
+    "min_absolute_lift": 0.55,
+    "optimized_score_max": 0.95,
+    "optimized_score_min": 0.9,
+    "optimizer_budget": {
+      "active_reservations": 0,
+      "exhausted": null,
+      "limits": {
+        "input_tokens": 1500000,
+        "output_tokens": 61440,
+        "requests": 240,
+        "usd": 3.0
+      },
+      "pricing": {
+        "input_per_million": 0.75,
+        "output_per_million": 4.5,
+        "source_url": "https://openai.com/api/pricing/"
+      },
+      "requests": 120,
+      "reservation_ledger_version": 2,
+      "reservations": [],
+      "reserved": {
+        "input_tokens": 0,
+        "output_tokens": 0,
+        "usd": 0.0
+      },
+      "single_attempt_transport_enforced": true,
+      "transport_attempts": 120,
+      "usage": {
+        "input_tokens": 42888,
+        "output_tokens": 1405,
+        "usd": 0.038487999999999994
+      }
+    },
+    "repeats": 3,
+    "total_usage": {
+      "input_tokens": 42888,
+      "output_tokens": 1405,
+      "requests_with_usage": 120,
+      "total_tokens": 44293
+    }
+  },
+  "tutorial": "docs/TUTORIAL_TICKET_ROUTING.md"
+}
diff --git a/decisions.md b/decisions.md
index 936b1ab0..92b4460b 100644
--- a/decisions.md
+++ b/decisions.md
@@ -13,7 +13,7 @@ not necessarily when it was made.
 | 2026-08-23 | Imp is not published to Hex; each release is a private immutable Git tag (`v0.3.x`) installed with `{:imp, github: "deepfates/imp", tag: ...}`. | `README.md` Install. `mix hex.build` refuses the Git ExMCP dependency, so `mix package.check` cannot establish Hex publishability; publication is an owner action. | Amended 2026-09-17: the repository is public, so the tags are public and need no credentials; whether to publish to Hex is still open. | ExMCP is consumable as a released Hex package and the owner chooses to publish. |
 | 2026-09-13 | An ACP tool kind is derived from the MCP `ToolAnnotations` the server declares, not from a host table keyed by tool name; a tool that declares no hint gets `nil` rather than a guess, and an explicit `:tool_kinds` entry still wins. | `lib/imp/acp/tool_kind.ex` moduledoc, `test/acp_options_tool_kinds_test.exs`. A name table goes stale the moment a server publishes a tool it does not name, and that is exactly when a permission mode that would have asked does not ask. | In force. | Does not retire. |
 | 2026-09-13 | On `session/load` and `session/resume` the session is installed with the `_meta` it was created with, not the `_meta` the request carried; both are passed to the program factory as `:meta` and `:requested_meta`. | `lib/imp/acp.ex` moduledoc, `test/acp_imp_acp_test.exs`. A session's history and transcript belong to the configuration that produced them, so a resume that silently adopts a different one replays one configuration's transcript as another's; only the factory knows which of its own `_meta` keys are identity-bearing, so it is given both and decides whether to refuse. | In force. | Does not retire. |
-| 2026-09-13 | `Imp.ACP` (the ACP session/program adapter) is part of Imp; the separate `imp_acp` package is retired with no compatibility shim and no second implementation. | Retired `imp_acp` README, `AGENTS.md`. One protocol stack, one owner of the ExMCP dependency. | In force on `main`; **not in `v0.3.2`** (the released tag predates absorption). | The "unreleased" caveat retires at the next tag. The absorption itself does not retire. |
+| 2026-09-13 | `Imp.ACP` (the ACP session/program adapter) is part of Imp; the separate `imp_acp` package is retired with no compatibility shim and no second implementation. | Retired `imp_acp` README, `AGENTS.md`. One protocol stack, one owner of the ExMCP dependency. | In force, and released in `v0.4.0`. | The "unreleased" caveat retired at `v0.4.0`. The absorption itself does not retire. |
 | 2026-09-13 | Absorption renames: `IMP_ACP_PATH` becomes `IMP_PATH` (copied workspace example), `IMP_ACP_ROOT` becomes `IMP_ROOT` (Haven preset), launcher is `examples/workspace_agent/scripts/workspace-agent-acp`. Module names `Imp.ACP`, `Imp.ACP.Host`, `Imp.ACP.MCP` and the `deepfates.com/imp-acp` wire metadata namespace keep their names. | Retired `imp_acp` README. Saved executable and environment references are updated explicitly, not by rewriting contact history. | In force. | Does not retire. |
 | 2026-09-13 | The workspace agent's session store keeps the `imp_acp/workspace_agent/sessions` directory name. | `examples/workspace_agent/README.md`. Preserves saved sessions; it is a storage location, not a dependency on the retired package. | In force. | Saved sessions are migrated, or the owner accepts losing restart continuity for them. |
 | 2026-09-13 | Ordinary Imp boot starts no protocol listener, subprocess or remote connection and does not start the ExMCP application. Protocol entry points (`Imp.ACP.*`, non-empty `Imp.MCP.connect/2`) start it explicitly. Releases using them declare `applications: [ex_mcp: :load]`. | `mix.exs` dependency comment, `docs/PRODUCTION_OPERATIONS.md` "Protocol runtime in releases", commit `728f8c77`. Prediction and optimizer processes must not open listeners or acquire protocol boot output. | In force. | Does not retire while Imp is a library inside a host application. |
diff --git a/docs/BENCHMARKS.md b/docs/BENCHMARKS.md
index 3875d0fe..b56e61d6 100644
--- a/docs/BENCHMARKS.md
+++ b/docs/BENCHMARKS.md
@@ -44,7 +44,7 @@ Also accepts `OPENROUTER_API_KEY` with `openai/gpt-5.4-mini`. Set
 
 - **Needs:** an OpenAI or OpenRouter key. No Python, no dataset download; the
   sixty tickets ship in the package at `priv/tutorial/support_tickets.json`.
-- **Time:** 8–9 seconds per repeat, three repeats by default.
+- **Time:** 8–13 seconds per repeat, three repeats by default.
 - **Cost:** about `$0.013` per repeat at `gpt-5.4-mini` prices; the script
   enforces a hard budget of `$1.00` per repeat and refuses to exceed it.
 - **What it measures:** held-out accuracy of a four-way enum router before and
diff --git a/docs/EVIDENCE.md b/docs/EVIDENCE.md
index 7fa17b58..735d06dd 100644
--- a/docs/EVIDENCE.md
+++ b/docs/EVIDENCE.md
@@ -45,9 +45,9 @@ compares and what it deliberately does not, are in
 Two held-out results, both small, both in RESULTS.md.
 
 Rows R1 and R2 are the ticket-routing tutorial: a zero-shot enum router scored
-0.30–0.50 on twenty held-out tickets and 0.95–1.00 after `LabeledFewShot(k: 8)`,
-over three live repeats at about a cent each. A stranger with an API key can run
-that command and get their own numbers. It is the only end-to-end effectiveness
+0.30–0.50 on twenty held-out tickets and 0.90–1.00 after `LabeledFewShot(k: 8)`,
+over three live repeats at about a cent each, in each of two runs a month apart.
+A stranger with an API key can run that command and get their own numbers. It is the only end-to-end effectiveness
 claim here that is reproducible from scratch.
 
 Row R6 is `examples/deployment/agent_optimization.exs`, the first ordinary Imp
diff --git a/docs/IMP_FOR_DSPY_USERS.md b/docs/IMP_FOR_DSPY_USERS.md
index 2c0259b3..bc7e7f8f 100644
--- a/docs/IMP_FOR_DSPY_USERS.md
+++ b/docs/IMP_FOR_DSPY_USERS.md
@@ -76,7 +76,7 @@ own held-out data.
 DSPy also has a larger Python integration ecosystem. Imp provides extension
 boundaries through `Imp.LM`, `Imp.Retrieve`, adapters, tools, and trainer
 clients; integrations written for Python do not automatically work on the
-BEAM. DSPy's Flex code optimizer is not included in `0.3.2`.
+BEAM. DSPy's Flex code optimizer is not included in `0.4.0`.
 
 ## Nearby Elixir work
 
@@ -90,7 +90,7 @@ lineage to a broad programming and optimization model built for OTP deployment.
 ## Coming from DSPy: the five-minute version
 
 ```elixir
-# pip install dspy            →  {:imp, github: "deepfates/imp", tag: "v0.3.2"}
+# pip install dspy            →  {:imp, github: "deepfates/imp", tag: "v0.4.0"}
 # dspy.configure(lm=lm)       →  lm = Imp.req_llm("openai:gpt-5.4-mini", api_key: ...)
 # dspy.Predict("q -> a")      →  program = Imp.predict("q -> a", lm: lm)
 # program(q="...")            →  {:ok, pred} = Imp.call(program, %{q: "..."})
diff --git a/docs/TUTORIAL_TICKET_ROUTING.md b/docs/TUTORIAL_TICKET_ROUTING.md
index 9268c3ab..219e76e9 100644
--- a/docs/TUTORIAL_TICKET_ROUTING.md
+++ b/docs/TUTORIAL_TICKET_ROUTING.md
@@ -4,10 +4,11 @@ Let's take the support-ticket router from the [README](../README.md) and do
 what you cannot do with a prompt string: score it on held-out data, improve it
 with an optimizer, and prove the improvement on tickets it has never seen.
 
-The zero-shot router scored **30–50%** on twenty held-out tickets; the
-optimized router scored **95–100%** on the same twenty in all three repeats.
-Each full experiment — baseline, optimization, and held-out evaluation — cost
-about **$0.013** and ran in **8–9 seconds** with `gpt-5.4-mini`.
+Across two runs a month apart, the zero-shot router scored **30–50%** on twenty
+held-out tickets; the optimized router scored **90–100%** on the same twenty in
+every repeat of both. Each full experiment — baseline, optimization, and
+held-out evaluation — used about 14,800 tokens and ran in **8–13 seconds** with
+`gpt-5.4-mini`, for an estimated **$0.013** at list prices.
 
 Those are rows R1 and R2 in
 [benchmarks/RESULTS.md](https://github.com/deepfates/imp/blob/main/benchmarks/RESULTS.md),
@@ -105,7 +106,8 @@ the misses are not random — they are the model guessing what squad names mean:
 
 It reads the tickets fine. It cannot know that atlas is the money squad. On a
 task this small the exact score moves a little between runs — our current
-repeats landed between 0.30 and 0.50 — but every run tells the same story.
+repeats landed between 0.30 and 0.50 across both runs — but every run tells the
+same story.
 
 ## Improve With Measured Lift
 
@@ -139,12 +141,14 @@ optimized = Imp.evaluate(compiled, testset, metric, max_concurrency: 8, timeout:
 #=> {0.3, 0.95}
 ```
 
-The three repeats measured 50% → 95%, 35% → 100%, and 30% → 95%: gains of
-45–65 points on held-out tickets. Each run used about 14,800 tokens, cost about
-$0.013, and finished in 8–9 seconds; all 120 evaluation calls completed without
-a row error and the cache was cleared before each repeat, so every call was
-live. Twenty rows move in 5-point steps, so trust the direction and the
-magnitude, not the endpoints. Rows R1 and R2 in
+The most recent three repeats measured 35% → 90%, 40% → 95%, and 30% → 95%:
+gains of 55–65 points on held-out tickets. Each run used about 14,800 tokens and
+finished in 8–13 seconds, an estimated $0.013 at list prices; all 120 evaluation
+calls completed without a row error and the cache was cleared before each
+repeat, so every call was live. An earlier run of the same command reached
+95–100% optimized with gains of 45–65 points, so treat 90% as the low end you
+should expect, not a regression. Twenty rows move in 5-point steps, so trust the
+direction and the magnitude, not the endpoints. Rows R1 and R2 in
 [benchmarks/RESULTS.md](https://github.com/deepfates/imp/blob/main/benchmarks/RESULTS.md).
 
 It is not a magic button. The remaining misses are genuinely marginal tickets
diff --git a/examples/deployment/README.md b/examples/deployment/README.md
index 2e46fc5e..e85dd44e 100644
--- a/examples/deployment/README.md
+++ b/examples/deployment/README.md
@@ -24,7 +24,7 @@ IMP_PATH=../.. mix deps.get
 IMP_PATH=../.. mix run --no-start run_workflow.exs
 ```
 
-A copied application can omit `IMP_PATH` and resolve the immutable `v0.3.2`
+A copied application can omit `IMP_PATH` and resolve the immutable `v0.4.0`
 Git dependency declared in `mix.exs`.
 
 The workflow:
diff --git a/examples/deployment/mix.exs b/examples/deployment/mix.exs
index 7a7d03a6..71fd8734 100644
--- a/examples/deployment/mix.exs
+++ b/examples/deployment/mix.exs
@@ -11,7 +11,7 @@ defmodule ImpDeployment.MixProject do
 
   defp deps do
     case System.get_env("IMP_PATH") do
-      nil -> [{:imp, github: "deepfates/imp", tag: "v0.3.2"}]
+      nil -> [{:imp, github: "deepfates/imp", tag: "v0.4.0"}]
       path -> [{:imp, path: path}]
     end
   end
diff --git a/livebooks/01_real_lm_front_door.livemd b/livebooks/01_real_lm_front_door.livemd
index 67104c05..05f96d42 100644
--- a/livebooks/01_real_lm_front_door.livemd
+++ b/livebooks/01_real_lm_front_door.livemd
@@ -30,7 +30,7 @@ if repo do
   Mix.install([{:imp, path: repo}], install_opts)
 else
   # Standalone notebook: install the tagged Git source release.
-  Mix.install([{:imp, github: "deepfates/imp", tag: "v0.3.2"}])
+  Mix.install([{:imp, github: "deepfates/imp", tag: "v0.4.0"}])
 end
 
 api_key = System.get_env("OPENAI_API_KEY") || System.get_env("LB_OPENAI_API_KEY")
diff --git a/livebooks/02_programming_not_prompting.livemd b/livebooks/02_programming_not_prompting.livemd
index e3ce8960..f1cc9442 100644
--- a/livebooks/02_programming_not_prompting.livemd
+++ b/livebooks/02_programming_not_prompting.livemd
@@ -30,7 +30,7 @@ if repo do
   Mix.install([{:imp, path: repo}], install_opts)
 else
   # Standalone notebook: install the tagged Git source release.
-  Mix.install([{:imp, github: "deepfates/imp", tag: "v0.3.2"}])
+  Mix.install([{:imp, github: "deepfates/imp", tag: "v0.4.0"}])
 end
 live_provider_enabled? = System.get_env("LIVE_PROVIDER") == "1"
 ```
diff --git a/livebooks/03_evaluate_and_optimize.livemd b/livebooks/03_evaluate_and_optimize.livemd
index c6b015b9..e3395ec0 100644
--- a/livebooks/03_evaluate_and_optimize.livemd
+++ b/livebooks/03_evaluate_and_optimize.livemd
@@ -30,7 +30,7 @@ if repo do
   Mix.install([{:imp, path: repo}], install_opts)
 else
   # Standalone notebook: install the tagged Git source release.
-  Mix.install([{:imp, github: "deepfates/imp", tag: "v0.3.2"}])
+  Mix.install([{:imp, github: "deepfates/imp", tag: "v0.4.0"}])
 end
 live_provider_enabled? = System.get_env("LIVE_PROVIDER") == "1"
 ```
diff --git a/livebooks/04_tools_agents_mcp_rlm.livemd b/livebooks/04_tools_agents_mcp_rlm.livemd
index 1414bcec..401bc5f2 100644
--- a/livebooks/04_tools_agents_mcp_rlm.livemd
+++ b/livebooks/04_tools_agents_mcp_rlm.livemd
@@ -30,7 +30,7 @@ if repo do
   Mix.install([{:imp, path: repo}], install_opts)
 else
   # Standalone notebook: install the tagged Git source release.
-  Mix.install([{:imp, github: "deepfates/imp", tag: "v0.3.2"}])
+  Mix.install([{:imp, github: "deepfates/imp", tag: "v0.4.0"}])
 end
 live_provider_enabled? = System.get_env("LIVE_PROVIDER") == "1"
 ```
diff --git a/livebooks/05_operate_and_live_checks.livemd b/livebooks/05_operate_and_live_checks.livemd
index 7c47384c..7b6d81af 100644
--- a/livebooks/05_operate_and_live_checks.livemd
+++ b/livebooks/05_operate_and_live_checks.livemd
@@ -30,7 +30,7 @@ if repo do
   Mix.install([{:imp, path: repo}], install_opts)
 else
   # Standalone notebook: install the tagged Git source release.
-  Mix.install([{:imp, github: "deepfates/imp", tag: "v0.3.2"}])
+  Mix.install([{:imp, github: "deepfates/imp", tag: "v0.4.0"}])
 end
 live_provider_enabled? = System.get_env("LIVE_PROVIDER") == "1"
 ```
diff --git a/mix.exs b/mix.exs
index a5484256..a1bef226 100644
--- a/mix.exs
+++ b/mix.exs
@@ -4,7 +4,7 @@ defmodule Imp.MixProject do
   def project do
     [
       app: :imp,
-      version: "0.3.2",
+      version: "0.4.0",
       elixir: "~> 1.19",
       name: "Imp",
       source_url: "https://github.com/deepfates/imp",
diff --git a/priv/public_api.json b/priv/public_api.json
index 3c575bbf..4cd98b42 100644
--- a/priv/public_api.json
+++ b/priv/public_api.json
@@ -9699,7 +9699,7 @@
       "types": []
     }
   ],
-  "package_version": "0.3.2",
+  "package_version": "0.4.0",
   "schema_version": 3,
   "scope": "Curated Imp API compiled from package-shipped sources. Exports come from non-hidden documentation entries; callbacks, types, and struct fields are category- and policy-gated."
 }
diff --git a/scripts/tutorial_ticket_routing_experiment.exs b/scripts/tutorial_ticket_routing_experiment.exs
index adf73dbc..99ceb6f9 100644
--- a/scripts/tutorial_ticket_routing_experiment.exs
+++ b/scripts/tutorial_ticket_routing_experiment.exs
@@ -159,9 +159,9 @@ defmodule TutorialTicketRoutingExperiment do
       "doc_claims_under_test" => %{
         "source" => "docs/TUTORIAL_TICKET_ROUTING.md",
         "baseline_repeat_range" => [0.30, 0.50],
-        "optimized_repeat_range" => [0.95, 1.0],
+        "optimized_repeat_range" => [0.90, 1.0],
         "cost_usd_per_repeat_approx" => 0.013,
-        "duration_seconds_range" => [8, 9]
+        "duration_seconds_range" => [8, 13]
       },
       "scope" => %{
         "claimed" =>
diff --git a/test/deployment_reference_test.exs b/test/deployment_reference_test.exs
index 4488ee0c..9a2486a7 100644
--- a/test/deployment_reference_test.exs
+++ b/test/deployment_reference_test.exs
@@ -454,7 +454,7 @@ defmodule DeploymentReferenceTest do
     readme = File.read!(Path.join(@example_root, "README.md"))
 
     assert mix_file =~ ~s(elixir: "~> 1.19")
-    assert mix_file =~ ~s({:imp, github: "deepfates/imp", tag: "v0.3.2"})
+    assert mix_file =~ ~s({:imp, github: "deepfates/imp", tag: "v0.4.0"})
     assert mix_file =~ "IMP_PATH"
     assert readme =~ "bounded supervised task"
     assert readme =~ "IMP_MODEL"
diff --git a/test/package_contract_test.exs b/test/package_contract_test.exs
index c67759a2..667a131b 100644
--- a/test/package_contract_test.exs
+++ b/test/package_contract_test.exs
@@ -106,11 +106,11 @@ defmodule PackageContractTest do
     assert_release_files(files)
   end
 
-  test "the 0.3.2 release surfaces agree" do
+  test "the 0.4.0 release surfaces agree" do
     version = Mix.Project.config()[:version]
     source_dependency = ~s({:imp, github: "deepfates/imp", tag: "v#{version}"})
 
-    assert version == "0.3.2"
+    assert version == "0.4.0"
     assert File.read!("RELEASE_NOTES.md") =~ "# Imp v#{version}"
     assert File.read!("CHANGELOG.md") =~ "## #{version}"
     assert File.read!("examples/deployment/mix.exs") =~ source_dependency