Value opencode Go plan requests; per-duration cache prices; usage under answers

- opencode Go reports plan requests as cost 0 with the usage in allowance;
  ModelConfig::cost values them at the model's list prices (charged amounts
  and free requests stay as reported). Rows booked at 0 are valued once on
  start (idempotent).
- opencode's price list now has per-duration cache prices for some models
  ({"30m": ..., "12h": ...}): the shortest duration is used, also in tiers.
  Saving the provider's models failed with "untagged enum ModelEntry".
- Each answer shows tokens sent/received and the cost of the turn's requests
  (usage.cost filled when messages are saved and loaded).
- The chat title keeps a minimum width in a crowded top bar.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

CHANGELOG.md

@@ -3,6 +3,23 @@
 All notable changes to CodeCraft. Versions follow [Semantic Versioning](https://semver.org/);
 until 1.0, minor versions may contain breaking changes.
 
+## [Unreleased]
+
+### Added
+
+- **Tokens and cost under every answer:** a small line under each answer of the model shows the
+  tokens sent/received and the cost of all requests of that turn.
+
+### Fixed
+
+- **Total cost of chats stopped growing (opencode Go):** opencode now reports plan requests as
+  "nothing charged" (cost 0, usage in `allowance`), and CodeCraft took the 0. Plan requests are
+  valued at the model's list prices again; earlier ones are corrected once on start.
+- **Saving opencode Go models failed after fetching prices** ("data did not match any variant of
+  untagged enum ModelEntry"): opencode lists cache prices per duration for some models
+  (`{"30m": …, "12h": …}`); the shortest duration is used now.
+- The chat title in the top bar no longer shrinks to a single letter when there are many chips.
+
 ## [v202611-00022] – 2026-11-11
 
 ### Added

core/src/config.rs

@@ -159,7 +159,13 @@ impl ModelConfig {
     /// Cost of one request now. Cache prices default to the input price.
     /// The provider's own figure wins when it reports one (opencode).
     pub fn cost(&self, u: &crate::types::Usage) -> f64 {
-        u.cost.unwrap_or_else(|| self.cost_at(u, chrono::Utc::now()))
+        match u.cost {
+            // A plan request (opencode Go): nothing charged, the usage counts toward the plan –
+            // its value at the model's list prices.
+            Some(c) if c == 0.0 && u.allowance.is_some_and(|a| a > 0.0) => self.cost_at(u, chrono::Utc::now()),
+            Some(c) => c,
+            None => self.cost_at(u, chrono::Utc::now()),
+        }
     }
 
     /// Cost of one request at a given time: context tiers and time factors applied.
@@ -187,6 +193,21 @@ mod price_tests {
         Usage { input_tokens: input, output_tokens: output, cache_read_tokens: read, ..Default::default() }
     }
 
+    #[test]
+    fn plan_requests_are_valued_at_list_prices() {
+        let m = ModelConfig { input_price: Some(1.0), output_price: Some(2.0), ..Default::default() };
+        let mut u = usage(1_000_000, 1_000_000, 0);
+        u.cost = Some(0.0);
+        u.allowance = Some(0.4);
+        assert_eq!(m.cost(&u), 3.0);
+        // Charged by the provider: that amount; free without a plan: free.
+        u.cost = Some(0.7);
+        assert_eq!(m.cost(&u), 0.7);
+        u.cost = Some(0.0);
+        u.allowance = None;
+        assert_eq!(m.cost(&u), 0.0);
+    }
+
     #[test]
     fn tiers_and_multipliers() {
         let m = ModelConfig {

server/src/api.rs

@@ -843,6 +843,19 @@ impl WindowQ {
     }
 }
 
+/// The cost of each answer for the chat view (usage.cost: charged, or valued at list prices).
+fn with_costs(app: &App, mut messages: Vec<crate::types::Message>) -> Vec<crate::types::Message> {
+    let registry = app.registry();
+    for m in &mut messages {
+        if let (Some(u), Some(spec)) = (m.usage.as_mut(), m.model.as_deref()) {
+            if let Ok((_, model)) = registry.resolve(spec) {
+                u.cost = Some(model.cost(u));
+            }
+        }
+    }
+    messages
+}
+
 async fn get_session(Extension(user): Extension<User>, State(app): State<AppState>, Path(id): Path<String>, Query(q): Query<WindowQ>) -> ApiResult {
     let s = readable(&app, &user, &id)?;
     let busy = app.is_busy(&id);
@@ -853,7 +866,7 @@ async fn get_session(Extension(user): Extension<User>, State(app): State<AppStat
     match q.window() {
         Some(w) => {
             let (messages, older, newer) = app.store.messages_window(&id, &w)?;
-            out["messages"] = json!(messages);
+            out["messages"] = json!(with_costs(&app, messages));
             out["has_older"] = json!(older);
             out["has_newer"] = json!(newer);
         }
@@ -867,7 +880,7 @@ async fn session_messages(Extension(user): Extension<User>, State(app): State<Ap
     readable(&app, &user, &id)?;
     let w = q.window().ok_or_else(|| ApiError(StatusCode::BAD_REQUEST, "before, after, around or limit is required".into()))?;
     let (messages, older, newer) = app.store.messages_window(&id, &w).map_err(|e| ApiError(StatusCode::NOT_FOUND, format!("{e:#}")))?;
-    Ok(Json(json!({"messages": messages, "has_older": older, "has_newer": newer})))
+    Ok(Json(json!({"messages": with_costs(&app, messages), "has_older": older, "has_newer": newer})))
 }
 
 /// Images, videos and audio files shown in a chat (all of it, with subagents).

server/src/app.rs

@@ -144,6 +144,11 @@ impl App {
                 Ok(n) => tracing::info!("moved {n} images of older chats to the image store"),
                 Err(e) => tracing::warn!("moving images to the image store: {e:#}"),
             }
+            match crate::budget::value_plan_requests(&self) {
+                Ok(0) => {}
+                Ok(n) => tracing::info!("valued {n} plan requests (opencode Go) at list prices"),
+                Err(e) => tracing::warn!("valuing plan requests: {e:#}"),
+            }
             match crate::images::backfill_copies(&self) {
                 Ok(0) => {}
                 Ok(n) => tracing::info!("kept copies of {n} generated images in the image store"),

server/src/budget.rs

@@ -361,6 +361,49 @@ fn opencode_when(w: &Value) -> Option<When> {
     Some((days, hours, from, until))
 }
 
+/// Plan requests (opencode Go) that were booked at 0: value them at the model's list
+/// prices of the time of the request. Idempotent – valued rows no longer match.
+pub fn value_plan_requests(app: &App) -> Result<usize> {
+    let mut n = 0;
+    for (row, ts, spec, usage) in app.store.plan_requests_without_cost()? {
+        let Ok((_, model)) = app.registry().resolve(&spec) else { continue };
+        let Some(at) = chrono::DateTime::from_timestamp_millis(ts) else { continue };
+        let cost = model.cost_at(&usage, at);
+        if cost > 0.0 {
+            app.store.set_usage_cost(row, cost)?;
+            n += 1;
+        }
+    }
+    Ok(n)
+}
+
+/// A price from opencode's list: a number, or one price per cache duration
+/// (`{"30m": 0.125, "12h": 0.25}`) – then the shortest duration (the default caching).
+fn price_of(v: &Value) -> Value {
+    match v {
+        Value::Number(_) => v.clone(),
+        Value::Object(o) => {
+            let secs = |k: &str| -> Option<f64> {
+                let (num, unit) = k.split_at(k.find(|c: char| c.is_ascii_alphabetic())?);
+                let n: f64 = num.parse().ok()?;
+                Some(n * match unit {
+                    "s" => 1.0,
+                    "m" => 60.0,
+                    "h" => 3600.0,
+                    "d" => 86400.0,
+                    _ => return None,
+                })
+            };
+            o.iter()
+                .filter(|(_, x)| x.is_number())
+                .min_by(|a, b| secs(a.0).unwrap_or(f64::MAX).total_cmp(&secs(b.0).unwrap_or(f64::MAX)))
+                .map(|(_, x)| x.clone())
+                .unwrap_or(Value::Null)
+        }
+        _ => Value::Null,
+    }
+}
+
 /// Model row from opencode's model list (prices per 1M tokens, tiers, time factors).
 fn opencode_model(m: &Value) -> Value {
     let cost = &m["cost"]["text"];
@@ -371,13 +414,13 @@ fn opencode_model(m: &Value) -> Value {
         for o in side["overrides"].as_array().into_iter().flatten() {
             let Some(gt) = o["when"]["context"]["gt"].as_u64() else { continue };
             let t = tiers.entry(gt).or_default();
-            t.insert(field.into(), o["price"].clone());
+            t.insert(field.into(), price_of(&o["price"]));
             if field == "input_price" {
-                if o["cache"]["read"].is_number() {
+                for (key, k) in [("read", "cache_read_price"), ("write", "cache_write_price")] {
-                    t.insert("cache_read_price".into(), o["cache"]["read"].clone());
+                    let p = price_of(&o["cache"][key]);
-                }
+                    if !p.is_null() {
-                if o["cache"]["write"].is_number() {
+                        t.insert(k.into(), p);
-                    t.insert("cache_write_price".into(), o["cache"]["write"].clone());
+                    }
                 }
             }
         }
@@ -412,10 +455,10 @@ fn opencode_model(m: &Value) -> Value {
         "max_input": m["limit"]["input"],
         "vision": images,
         "reasoning": m["capabilities"]["reasoning"]["supported"].as_bool().unwrap_or(false),
-        "input_price": input["price"],
+        "input_price": price_of(&input["price"]),
-        "output_price": output["price"],
+        "output_price": price_of(&output["price"]),
-        "cache_read_price": input["cache"]["read"],
+        "cache_read_price": price_of(&input["cache"]["read"]),
-        "cache_write_price": input["cache"]["write"],
+        "cache_write_price": price_of(&input["cache"]["write"]),
         "price_tiers": price_tiers,
         "price_multipliers": price_multipliers,
         "notes": notes,
@@ -481,6 +524,25 @@ pub async fn discover_prices(app: &App, provider: &str) -> Result<Value> {
 
 #[cfg(test)]
 mod tests {
+    #[test]
+    fn prices_per_cache_duration() {
+        use super::{opencode_model, price_of};
+        use serde_json::{Value, json};
+        assert_eq!(price_of(&json!(0.5)), json!(0.5));
+        assert_eq!(price_of(&json!({"12h": 0.25, "30m": 0.125})), json!(0.125));
+        assert_eq!(price_of(&json!({"1h": 2.0, "5m": 1.25})), json!(1.25));
+        assert_eq!(price_of(&Value::Null), Value::Null);
+        // A model of opencode's list with per-duration cache write prices.
+        let m = json!({"id": "gpt-6-luna", "limit": {"context": 1050000, "input": 922000, "output": 128000},
+            "cost": {"text": {"input": {"price": 0.1, "cache": {"read": 0.01, "write": {"30m": 0.125, "12h": 0.25}},
+                "overrides": [{"when": {"context": {"gt": 272000}}, "price": 0.2, "cache": {"read": 0.02, "write": {"30m": 0.25, "12h": 0.5}}}]},
+                "output": {"price": 0.5, "overrides": [{"when": {"context": {"gt": 272000}}, "price": 0.75}]}}}});
+        let row = opencode_model(&m);
+        assert_eq!(row["cache_write_price"], json!(0.125));
+        assert_eq!(row["price_tiers"][0]["cache_write_price"], json!(0.25));
+        assert_eq!(row["max_input"], json!(922000));
+    }
+
     #[test]
     fn opencode_cost_conversion() {
         let m: serde_json::Value = serde_json::from_str(r#"{"id":"grok","limit":{"context":500000,"output":8000},

server/src/runner.rs

@@ -1113,7 +1113,8 @@ async fn stream_step(
         let mut message = Message::new(Role::Assistant, parts);
         message.id = msg_id;
         message.model = Some(model_spec.to_string());
-        message.usage = Some(usage.clone());
+        // With its cost (charged, or valued at list prices) for the line under the answer.
+        message.usage = Some(Usage { cost: Some(req.model.cost(&usage)), ..usage.clone() });
         if !message.parts.is_empty() {
             app.store.add_message(&session.id, &message)?;
             app.emit(Event::MessageAdded { session_id: session.id.clone(), message: message.clone() });

server/src/store.rs

@@ -653,6 +653,39 @@ impl Store {
         Ok(json!({"total": total, "by_model": by_model, "by_agent": by_agent, "first": first, "last": last}))
     }
 
+    /// Plan requests booked at 0 (opencode Go reports nothing charged, only `allowance`):
+    /// (row id, ts, "provider/model", usage).
+    pub fn plan_requests_without_cost(&self) -> Result<Vec<(i64, i64, String, crate::types::Usage)>> {
+        let conn = self.conn.lock().unwrap();
+        let mut stmt = conn.prepare(
+            "SELECT id, ts, provider || '/' || model, input_tokens, output_tokens, cache_read_tokens, cache_write_tokens, allowance
+             FROM usage WHERE cost = 0 AND allowance > 0",
+        )?;
+        let rows = stmt
+            .query_map([], |r| {
+                Ok((
+                    r.get(0)?,
+                    r.get(1)?,
+                    r.get(2)?,
+                    crate::types::Usage {
+                        input_tokens: r.get::<_, i64>(3)? as u64,
+                        output_tokens: r.get::<_, i64>(4)? as u64,
+                        cache_read_tokens: r.get::<_, i64>(5)? as u64,
+                        cache_write_tokens: r.get::<_, i64>(6)? as u64,
+                        allowance: r.get(7)?,
+                        ..Default::default()
+                    },
+                ))
+            })?
+            .collect::<rusqlite::Result<Vec<_>>>()?;
+        Ok(rows)
+    }
+
+    pub fn set_usage_cost(&self, row: i64, cost: f64) -> Result<()> {
+        self.conn.lock().unwrap().execute("UPDATE usage SET cost = ?2 WHERE id = ?1", params![row, cost])?;
+        Ok(())
+    }
+
     /// Global statistics in [from, to) (ms). `tz_offset_min` shifts day/hour buckets
     /// into the viewer's local time.
     pub fn global_stats(&self, from: i64, to: i64, tz_offset_min: i64, user: Option<&str>) -> Result<Value> {

web/src/components/ChatView.tsx

@@ -245,8 +245,10 @@ const MessageList = memo(function MessageList({
   baseDir,
   onRewind,
   highlight,
+  currency,
 }: {
   messages: Message[];
+  currency: string;
   results: Map<string, ToolResult>;
   busy: boolean;
   canEdit: boolean;
@@ -280,13 +282,52 @@ const MessageList = memo(function MessageList({
         ));
         // Finished turns with tool steps fold into one line (the answer stays visible).
         const steps = b.messages.flatMap((m) => m.parts.filter((p) => p.type === "tool_call")) as ToolCall[];
-        if (i === lastWork || steps.length < 2) return <div key={key}>{items}</div>;
+        const usage = !(busy && i === lastWork) && <TurnUsage messages={b.messages} currency={currency} />;
-        return <WorkBlock key={key} messages={b.messages} steps={steps} open={!!highlight && b.messages.some((m) => m.id === highlight)} baseDir={baseDir}>{items}</WorkBlock>;
+        if (i === lastWork || steps.length < 2)
+          return (
+            <div key={key}>
+              {items}
+              {usage}
+            </div>
+          );
+        return (
+          <div key={key}>
+            <WorkBlock messages={b.messages} steps={steps} open={!!highlight && b.messages.some((m) => m.id === highlight)} baseDir={baseDir}>
+              {items}
+            </WorkBlock>
+            {usage}
+          </div>
+        );
       })}
     </>
   );
 });
 
+/** Tokens (in/out) and cost of one answer: all requests of the turn together. */
+function TurnUsage({ messages, currency }: { messages: Message[]; currency: string }) {
+  const t = useT();
+  let inp = 0;
+  let out = 0;
+  let cost = 0;
+  let requests = 0;
+  for (const m of messages) {
+    const u = m.role === "assistant" ? m.usage : undefined;
+    if (!u) continue;
+    requests++;
+    inp += u.input_tokens + u.cache_read_tokens + u.cache_write_tokens;
+    out += u.output_tokens;
+    cost += u.cost ?? 0;
+  }
+  if (!requests) return null;
+  return (
+    <div className="turn-usage" title={t("chat.turnUsageHint", { n: requests })}>
+      {fmtCompact(inp)}
+      <Icon name="arrow-up" className="tiny" /> {fmtCompact(out)}
+      <Icon name="arrow-down" className="tiny" /> · {fmtCost(cost, currency)}
+    </div>
+  );
+}
+
 /** A finished turn: "42 steps · 3 files" and the answer; the steps on click. */
 function WorkBlock({ messages, steps, open: forced, baseDir, children }: { messages: Message[]; steps: ToolCall[]; open: boolean; baseDir: string; children: ReactNode }) {
   const t = useT();
@@ -825,6 +866,7 @@ export function ChatView({ sessionId, onMenu, onClosePane }: { sessionId: string
             baseDir={session.cwd}
             onRewind={onRewind}
             highlight={highlight}
+            currency={stats?.currency ?? "USD"}
           />
           {win?.newer && (
             <button className="back-to-latest" onClick={() => void showLatest()}>

web/src/i18n/de.ts

@@ -367,6 +367,7 @@ export const de: Messages = {
   "media.prev": "Vorheriges",
   "media.next": "Nächstes",
   "media.back": "Zurück zur Übersicht",
+  "chat.turnUsageHint": "Alle {n} Anfragen dieser Antwort (Tokens gesendet/empfangen, Kosten)",
   "chat.reviewRound": "Review {round}: Änderungen nötig",
   "agents.refine": "Review-Schleife",
   "agents.refineRounds": "Review-Runden pro Aufgabe",

web/src/i18n/en.ts

@@ -366,6 +366,7 @@ export const en = {
   "media.prev": "Previous",
   "media.next": "Next",
   "media.back": "Back to the overview",
+  "chat.turnUsageHint": "All {n} requests of this answer (tokens sent/received, cost)",
   "chat.reviewRound": "Review {round}: changes requested",
   "agents.refine": "Review loop",
   "agents.refineRounds": "Review rounds per task",

web/src/styles.css

@@ -456,7 +456,8 @@ button.session-group:hover {
 }
 .topbar .grow {
   flex: 1;
-  min-width: 0;
+  /* The title stays readable: with many chips they wrap into a second row instead. */
+  min-width: min(160px, 40vw);
 }
 .topbar h1 {
   font-size: 15px;
@@ -4398,3 +4399,14 @@ span.chat-tab > button {
   text-overflow: ellipsis;
   white-space: nowrap;
 }
+
+/* Tokens and cost under each answer. */
+.turn-usage {
+  display: flex;
+  align-items: center;
+  gap: 2px;
+  margin: -6px 0 12px;
+  font-size: 11px;
+  color: var(--muted);
+  font-variant-numeric: tabular-nums;
+}

web/src/types.ts

@@ -31,6 +31,8 @@ export interface Usage {
   output_tokens: number;
   cache_read_tokens: number;
   cache_write_tokens: number;
+  /** Cost of the request (charged, or valued at list prices for plan requests). */
+  cost?: number;
 }
 
 export interface Message {

web/src/whatsnew.ts

@@ -3,6 +3,10 @@
  * under "unreleased"; scripts/release.sh renames the key to the release version.
  */
 export const WHATS_NEW: Record<string, { en: string[]; de: string[] }> = {
+  unreleased: {
+    en: ["Tokens and cost under every answer; the total cost of chats with opencode Go is shown correctly again"],
+    de: ["Tokens und Kosten unter jeder Antwort; der Gesamtpreis von Chats mit opencode Go stimmt wieder"],
+  },
   "202611-00022": {
     en: ["New theme “Tintenblau” (Settings → General → Theme)"],
     de: ["Neues Theme „Tintenblau“ (Einstellungen → Allgemein → Theme)"],