# v0.5.20 (2026-07-07)

## Features
- **Thinking**: per-model thinking level picker on provider page — appends `(level)` suffix to copied model names for forced reasoning effort across all formats (openai, claude, gemini, deepseek, kimi, qwen, zai, minimax, hunyuan, step)
- **RTK**: add JS-native git-log filter (#2423)
- **Caveman**: add targeted upstream-aligned style rules (#2424)
- **i18n**: add Farsi (fa) language support (#2385)

## Fixes
- **Thinking**: strip `(level)` suffix from upstream `body.model` so providers no longer reject requests
- **Translator**: preserve developer instructions in openai-responses conversion (#2434)
- **count_tokens**: count structured Anthropic blocks (#2419)
- **Volcengine-ark**: clamp GLM-5 max_tokens to model output ceiling (#2428)
- **Kimi**: normalize reasoning_effort to backend enum (#2427)
- **Claude**: reconcile max_tokens vs thinking budget and lift per-model ceiling (#2381)
- **Kiro**: deliver system prompt natively, add Opus 4.5/4.7/4.8, tolerate dash version ids (#2366)
- **Headroom**: proxy dashboard through app (#2372)
- **MITM**: recover from stale lock file on server start

open-sse/config/providerModels.js

@@ -63,14 +63,19 @@ export function getModelType(aliasOrId, modelId) {
 }
 
 export function getModelUpstreamId(aliasOrId, modelId) {
+  // Split off thinking suffix "(level)" so lookup hits the base id; re-append it to
+  // the result so downstream applyThinking still sees the suffix (body.model is stripped separately).
+  const sufMatch = typeof modelId === "string" ? modelId.match(/\([^()]+\)\s*$/) : null;
+  const suffix = sufMatch ? sufMatch[0] : "";
+  const baseId = suffix ? modelId.slice(0, sufMatch.index).trim() : modelId;
   const models = PROVIDER_MODELS[aliasOrId];
-  const found = findModel(models, modelId, aliasOrId);
-  if (found?.upstreamModelId) return found.upstreamModelId;
-  if (found?.id) return found.id;
-  if (aliasOrId === "cx" && typeof modelId === "string" && modelId.endsWith(CODEX_REVIEW_SUFFIX)) {
-    return modelId.slice(0, -CODEX_REVIEW_SUFFIX.length);
+  const found = findModel(models, baseId, aliasOrId);
+  if (found?.upstreamModelId) return found.upstreamModelId + suffix;
+  if (found?.id) return found.id + suffix;
+  if (aliasOrId === "cx" && typeof baseId === "string" && baseId.endsWith(CODEX_REVIEW_SUFFIX)) {
+    return baseId.slice(0, -CODEX_REVIEW_SUFFIX.length) + suffix;
   }
-  return modelId;
+  return baseId + suffix;
 }
 
 export function getModelQuotaFamily(aliasOrId, modelId) {

open-sse/handlers/chatCore.js

@@ -1,5 +1,6 @@
 import { detectFormat, getTargetFormat, resolveTransport } from "../services/provider.js";
 import { translateRequest } from "../translator/index.js";
+import { stripThinkingSuffix } from "../translator/concerns/thinkingUnified.js";
 import { FORMATS } from "../translator/formats.js";
 import { normalizeClaudePassthrough } from "../translator/formats/claude.js";
 import { COLORS } from "../utils/stream.js";
@@ -123,9 +124,9 @@ export async function handleChatCore({ body, modelInfo, credentials, log, onCred
   let toolNameMap;
   if (passthrough) {
     log?.debug?.("PASSTHROUGH", `${clientTool} → ${provider} | native lossless`);
-    translatedBody = { ...body, model: upstreamModel };
+    translatedBody = { ...body, model: stripThinkingSuffix(upstreamModel) };
     // Normalize newer Cowork/CC beta shapes (adaptive thinking, mid-conversation system) the API rejects
-    if (clientTool === "claude") normalizeClaudePassthrough(translatedBody, upstreamModel);
+    if (clientTool === "claude") normalizeClaudePassthrough(translatedBody, translatedBody.model);
   } else {
     translatedBody = translateRequest(sourceFormat, targetFormat, upstreamModel, body, stream, credentials, provider, reqLogger, stripList, connectionId, clientTool);
     if (!translatedBody) {
@@ -134,7 +135,7 @@ export async function handleChatCore({ body, modelInfo, credentials, log, onCred
     }
     toolNameMap = translatedBody._toolNameMap;
     delete translatedBody._toolNameMap;
-    translatedBody.model = upstreamModel;
+    translatedBody.model = stripThinkingSuffix(upstreamModel);
   }
 
   // Dedupe duplicate built-in tools when equivalent MCP tools are present (Claude clients only).

open-sse/providers/registry/claude.js

@@ -60,12 +60,10 @@ export default {
     },
   },
   models: [
+    { id: "claude-fable-5", name: "Claude Fable 5" },
+    { id: "claude-sonnet-5", name: "Claude Sonnet 5" },
     { id: "claude-opus-4-8", name: "Claude Opus 4.8" },
     { id: "claude-opus-4-7", name: "Claude Opus 4.7" },
-    { id: "claude-opus-4-6", name: "Claude Opus 4.6" },
-    { id: "claude-sonnet-4-6", name: "Claude Sonnet 4.6" },
-    { id: "claude-opus-4-5-20251101", name: "Claude 4.5 Opus" },
-    { id: "claude-sonnet-4-5-20250929", name: "Claude 4.5 Sonnet" },
     { id: "claude-haiku-4-5-20251001", name: "Claude 4.5 Haiku" },
   ],
   oauth: {

open-sse/providers/thinkingLevels.js

@@ -0,0 +1,46 @@
+// Resolve valid thinking levels per model — drives UI level picker (suffix "model(level)").
+// Reuses capabilities.js (thinkingFormat/canDisable) so this file only maps format→levels (DRY).
+import { getCapabilitiesForModel } from "./capabilities.js";
+import { matchPattern } from "./pricing.js";
+
+// Shared level sets (deduped) — verified against provider docs + wire in thinkingUnified.applyFormat.
+const L = {
+  base: ["none", "low", "medium", "high"],                          // qwen, step, hunyuan, gemini-budget
+  onOff: ["none", "thinking"],                                      // zai (binary), minimax (adaptive)
+  openai: ["none", "minimal", "low", "medium", "high", "xhigh"],    // GPT-5.x / o-series (no "max")
+  levelMax: ["none", "low", "medium", "high", "max"],               // claude-adaptive, kimi
+  budgetX: ["none", "low", "medium", "high", "xhigh", "max"],       // claude-budget
+  gemini: ["minimal", "low", "medium", "high"],                     // gemini-3 thinkingLevel (no disable)
+  hiMax: ["none", "high", "max"],                                   // deepseek (low/med→high, xhigh→max)
+};
+
+// thinkingFormat → valid selectable levels (source of truth for UI options).
+const FORMAT_LEVELS = {
+  openai: L.openai,
+  "claude-adaptive": L.levelMax,
+  "claude-budget": L.budgetX,
+  "gemini-level": L.gemini,
+  "gemini-budget": L.base,
+  zai: L.onOff,
+  qwen: L.base,
+  kimi: L.levelMax,
+  deepseek: L.hiMax,
+  minimax: L.onOff,
+  hunyuan: L.base,
+  step: L.base,
+};
+
+// Model-name pattern overrides (glob, first match wins) — more precise than format default.
+const PATTERN_THINKING = [
+  { pattern: "*codex*", levels: ["low", "medium", "high", "xhigh"] }, // codex cannot disable thinking
+];
+
+// Returns valid thinking levels for a model, or null when the model has no reasoning.
+export function getThinkingLevels(provider, model) {
+  const caps = getCapabilitiesForModel(provider, model);
+  if (!caps.reasoning) return null;
+  const hit = PATTERN_THINKING.find((p) => matchPattern(p.pattern, model));
+  let levels = hit?.levels || FORMAT_LEVELS[caps.thinkingFormat] || L.base;
+  if (caps.thinkingCanDisable === false) levels = levels.filter((l) => l !== "none");
+  return levels;
+}

open-sse/translator/concerns/thinkingUnified.js

@@ -20,6 +20,13 @@ const FORMAT_TO_NATIVE = {
   kiro: "kiro",
 };
 
+// Strip a trailing thinking suffix "model(value)" → "model" (no-op when absent).
+export function stripThinkingSuffix(model) {
+  if (typeof model !== "string") return model;
+  const m = model.match(/^(.*)\([^()]+\)\s*$/);
+  return m ? m[1].trim() : model;
+}
+
 // Parse model-name suffix "model(value)" → { cleanModel, override }.
 // value: level name (high) | number (8192) | auto | none. null override when absent.
 export function parseSuffix(model) {