free-coding-models 0.5.49 → 0.5.51

package/README.md

@@ -539,7 +539,7 @@ Cerebras free-tier API has a strict **~8k total token limit** (prompt + tools +
 
 ### Full documentation
 
-See [`pi-extension/README.md`](./pi-extension/README.md) for the complete architecture diagram, config format, and troubleshooting guide.
+The adapter lives in [`packages/fcm-pi`](./packages/fcm-pi); `pi-extension/` is kept as a thin compat wrapper so existing local-path installs keep loading. See [`packages/fcm-pi/README.md`](./packages/fcm-pi/README.md) for the complete architecture, config format, and the one-time `packages/` self-link setup.
 
 ---
 
@@ -551,10 +551,12 @@ See [`pi-extension/README.md`](./pi-extension/README.md) for the complete archit
 
 ```bash
 mkdir -p ~/.config/opencode/plugins
-ln -sf /Users/vava/Documents/GitHub/free-coding-models/opencode-plugin/index.js \
+ln -sf /Users/vava/Documents/GitHub/free-coding-models/packages/fcm-opencode/index.js \
   ~/.config/opencode/plugins/fcm-opencode.js
 ```
 
+(`opencode-plugin/` at the repo root is kept as a thin compat wrapper, so existing symlinks keep working.)
+
 ### Commands
 
 | Command | Description |
@@ -568,7 +570,31 @@ ln -sf /Users/vava/Documents/GitHub/free-coding-models/opencode-plugin/index.js
 
 Startup is intentionally light: fresh cache first, daemon second, **no direct scan** unless you run `/fcm`.
 
-See [`opencode-plugin/README.md`](./opencode-plugin/README.md) for details and limitations.
+See [`packages/fcm-opencode/README.md`](./packages/fcm-opencode/README.md) for details and limitations.
+
+---
+
+## 🧩 Agent extensions architecture
+
+Both adapters share one core so scan/rank/cache/daemon/API-key/provider logic lives in exactly one place:
+
+```
+packages/
+├── fcm-agent-core/   ← shared core (scan, rank, cache, daemon, keys, provider descriptors; no rendering)
+├── fcm-pi/           ← Pi adapter (hooks, commands, status-bar renderer, ~/.pi/agent disk writer)
+└── fcm-opencode/     ← OpenCode adapter (config mutation, commands, toasts, shell.env)
+```
+
+- The core emits **structured progress events**; each adapter renders them its own way (Pi status bar, OpenCode toast).
+- API keys are never inlined into OpenCode config — they are referenced via `{env:FCM_<PROVIDER>_API_KEY}`.
+- The cross-tool cache means a scan done in Pi benefits OpenCode (and vice-versa).
+
+> The `packages/` tree needs a one-time self-link so `free-coding-models` resolves by name during local-path use:
+> ```bash
+> cd packages && mkdir -p node_modules && ln -s ../../ node_modules/free-coding-models
+> ```
+
+See [`packages/fcm-agent-core/README.md`](./packages/fcm-agent-core/README.md) for the public API and rationale. Tracked in [`.kandown/tasks/t2.md`](./.kandown/tasks/t2.md).
 
 ---
 
@@ -783,6 +809,7 @@ Telemetry is enabled by default and can be disabled with any of the following:
     <td align="center" width="120"><a href="https://github.com/serajbaltu"><img src="https://avatars.githubusercontent.com/u/90699173?v=4&s=80" width="80" height="80" style="border-radius:50%" alt="serajbaltu"></a></td>
     <td align="center" width="120"><a href="https://github.com/stgreenb"><img src="https://avatars.githubusercontent.com/u/18483964?v=4&s=80" width="80" height="80" style="border-radius:50%" alt="stgreenb"></a></td>
     <td align="center" width="120"><a href="https://github.com/MoriDanWork"><img src="https://avatars.githubusercontent.com/u/55363096?v=4&s=80" width="80" height="80" style="border-radius:50%" alt="MoriDanWork"></a></td>
+    <td align="center" width="120"><a href="https://github.com/fan92rus"><img src="https://avatars.githubusercontent.com/u/13201333?v=4&s=80" width="80" height="80" style="border-radius:50%" alt="fan92rus"></a></td>
   </tr>
   <tr>
     <td align="center"><a href="https://github.com/vava-nessa"><sub><b>vava-nessa</b></sub></a></td>
@@ -794,6 +821,10 @@ Telemetry is enabled by default and can be disabled with any of the following:
     <td align="center"><a href="https://github.com/serajbaltu"><sub><b>serajbaltu</b></sub></a></td>
     <td align="center"><a href="https://github.com/stgreenb"><sub><b>stgreenb</b></sub></a></td>
     <td align="center"><a href="https://github.com/MoriDanWork"><sub><b>MoriDanWork</b></sub></a></td>
+    <td align="center"><a href="https://github.com/fan92rus"><sub><b>fan92rus</b></sub></a></td>
+  </tr>
+  <tr>
+    <td align="center" colspan="10"><sub>🛡️ <b>fan92rus</b> — Windows path traversal fix (<code>path.sep</code>)</sub></td>
   </tr>
 </table>
 

package/changelog/v0.5.48.md

@@ -0,0 +1,25 @@
+# Changelog v0.5.48 - 2026-07-15
+
+### Added
+- **Quota column across TUI and Web Dashboard** — users can now see remaining free quota directly in the model table when a provider exposes reliable telemetry. The column shows a percentage for monitorable providers, `N/A` for providers where quota cannot be monitored safely, and `-` when the provider is not configured.
+- **Groq, Cerebras, and OpenRouter quota monitoring** — Groq and Cerebras are parsed from official rate-limit response headers, while OpenRouter continues to use its official key endpoint. Cerebras support includes both daily request headers and token-per-minute headers.
+- **Shared `fcm-agent-core` package** (`packages/fcm-agent-core/`) — all scan, ranking, cache, daemon, API-key, and provider-descriptor logic now lives in one reusable core consumed by both the Pi and OpenCode adapters. The core emits *structured* progress events (`{ phase, percent, completed, total, activeModels, message }`) and contains **no rendering and no chalk** — each adapter renders its own way. Public API: `scanBestFcmModel`, `directScan`, `rankModels`, `createCacheStore`, `buildPiProviderDescriptor`, `buildOpenCodeProviderDescriptor`, `buildSmartRouterDescriptor`, and more.
+- **Canonical `fcm-pi` adapter** (`packages/fcm-pi/`) and **canonical `fcm-opencode` adapter** (`packages/fcm-opencode/`) — the Pi and OpenCode host adapters, now containing only host-specific concerns (Pi status-bar renderer + `~/.pi/agent` disk writer; OpenCode config mutation + toasts + `shell.env`).
+- **21 new unit tests** for the core (`test/fcm-agent-core.test.js`): SWE parsing, composite scoring + ranking, context-window parsing and the 16k agent safety floor, base-URL normalization, provider-ID prefixing, the Pi/OpenCode/router descriptor shapes, a **no-API-key-leak assertion** for OpenCode (keys are referenced via `{env:FCM_*_API_KEY}`), and the namespaced cache store (round-trip, TTL expiry, legacy fallback, context re-filtering).
+- **Quota parser tests** for Groq and Cerebras header variants, covering request/day and token/minute quota shapes.
+
+### Changed
+- **Three-package layout under `packages/`.** The Pi adapter, OpenCode adapter, and shared core are now cleanly separated. Provider config shapes that were previously duplicated (the Pi provider shape existed both in runtime registration and in disk writes) are now centralized in `buildPiProviderDescriptor` / `buildOpenCodeProviderDescriptor`.
+- **Cross-tool cache.** Cache is now a shared `createCacheStore({ filePath, legacyPaths, ttlMs })`; the OpenCode adapter falls back to the Pi cache location, so a scan done in one tool benefits the other.
+- **Rendering decoupled from scanning.** The direct scanner no longer imports chalk or builds ANSI. The branded `> free-coding-models` Pi footer badge (exact TUI header green/white-on-black) now lives in the Pi renderer (`packages/fcm-pi/lib/pi-progress-renderer.js`), where it belongs.
+- **`pi-extension/` and `opencode-plugin/` are now thin compat wrappers** that re-export from `packages/fcm-pi` and `packages/fcm-opencode` respectively, so existing local-path/symlink installs keep loading unchanged. The moved `lib/` code was deleted (now in the core + adapters).
+- README updated with the new **Agent extensions architecture** section and corrected package links.
+- `pnpm test` now runs `test/fcm-agent-core.test.js` alongside `test/test.js`.
+- Provider quota capability metadata now only marks providers as percent-monitorable when they have verified telemetry, preventing misleading quota values for NVIDIA, OpenCode Zen, Routeway, and other unknown providers.
+
+### Fixed
+- Eliminates the provider-shape duplication flagged in `.kandown/tasks/t2.md`: Pi runtime registration and `~/.pi/agent/models.json` disk writes now share one descriptor builder, so they can never drift apart.
+
+### Notes
+- The adapters remain **BETA** and are still installed via local path / symlink (not via the npm `free-coding-models` package). Local use needs a one-time self-link: `cd packages && mkdir -p node_modules && ln -s ../../ node_modules/free-coding-models` (documented in `packages/fcm-agent-core/README.md`; same trick the legacy `pi-extension/` already used).
+- Task `t2` (*« Mutualiser FCM-Pi et créer fcm-opencode »*) is now **Done** — all acceptance criteria and the definition of done are met. Deliberately deferred to v1.1: npm publishing of the sub-packages, OpenCode custom `tool` definitions (v1 uses command hooks), TypeScript build for `fcm-opencode` (kept JS + JSDoc), and unifying `src/core/endpoint-installer.js` onto the shared descriptor builders.

package/changelog/v0.5.50.md

@@ -0,0 +1,21 @@
+# Changelog v0.5.50 - 2026-07-18
+
+### Changed
+
+- **Full model catalog audit** — verified all 20 providers against their official docs and APIs (2026-07-18)
+- **Auto-deprecation support** — models with a `deprecatedAfter` date in the tuple are now automatically filtered from `MODELS` after that date, without requiring a code change
+- **Cerebras** — `zai-glm-4.7` marked as `⚠️ DEPRECATED shutdown 2026-08-17` per official Cerebras notice; will disappear automatically after that date
+
+### Fixed
+
+- **NVIDIA NIM** — removed `magistral-small-2506` (not found on NIM) and `qwen3-coder-480b` (deprecated free endpoint); fixed context windows: `mistral-medium-3.5-128b` 128k→256k, `nemotron-3-nano-omni` 128k→256k, `mistral-nemotron` -→128k; added `deepseek-r1` at A tier
+- **Codestral** — fixed `codestral-2508` context window 128k→256k
+- **OpenRouter** — removed `gpt-oss-120b:free` and `gpt-oss-20b:free` (no longer available on free tier)
+- **Ollama Cloud** — full catalog refresh: 19 outdated models removed, 10 new models added (kimi-k2.6, kimi-k2.7-code, kimi-k2.5, deepseek-v4-flash, deepseek-v4-pro, glm-5.1, glm-5.2, minimax-m2.7, and more); catalog now verified against official ollama.com/search?c=cloud
+- **Novita AI** — removed stale models (qwen3.6-plus, nex-agi/nex-n2-pro, minimax/m2-her); added tencent/hy3
+- **LLM7** — removed devstral-small-2:24b; added minimax-m2.7 (S+) and gpt-oss:20b (A+)
+- **OpenCode Zen** — added hy3-free (Tencent Hy3 Free)
+
+### Confirmed (no changes)
+
+- groq, googleai, github-models, mistral, cloudflare, sambanova, ovhcloud, zai, scaleway, qwen, kilo, routeway — all models verified current

package/changelog/v0.5.51.md

@@ -0,0 +1,15 @@
+# Changelog v0.5.51 - 2026-07-25
+
+### Fixed
+- 🛡️ **Windows path traversal protection** — `serveWebStaticFile()` in `src/core/router-daemon.js` now uses `path.sep` instead of a hardcoded `/` for the path traversal guard. On Windows, `path.resolve()` returns backslash separators, which made the previous `candidate.startsWith(WEB_DIST_DIR + '/')` check always fail and serve `path_traversal_blocked` for every root request. On Linux/macOS, `path.sep === '/'` so behaviour is unchanged. *(Thanks @fan92rus! First contribution 🎉)*
+
+### Changed
+- 📦 **Deps:** bump `@tanstack/react-virtual` from `3.14.3` → `3.14.5`
+- 📦 **DevDeps:** bump `@vitejs/plugin-react` from `6.0.2` → `6.0.3`
+
+### Maintenance
+- 🧹 **kandown housekeeping:** drop stale `AGENT.md` / `AGENT_KANDOWN.md` copies, repair install, refresh agent ref
+- 🧹 **fcm-agent fixes:** `direct-scanner.js`, `ranker.js`, `scan-orchestrator.js`, `pi-progress-renderer.js`, and `extensions/index.js` refinements (312+/46-)
+
+### Contributors
+- **@fan92rus** — Windows path traversal fix (`path.sep`). Welcome aboard! 🛡️🎉

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "free-coding-models",
-  "version": "0.5.49",
+  "version": "0.5.51",
   "description": "Find the fastest coding LLM models in seconds — ping free models from multiple providers, pick the best one for OpenCode, Cursor, or any AI coding assistant.",
   "keywords": [
     "nvidia",

package/sources.js

@@ -51,16 +51,15 @@ export const nvidiaNim = [
   // ── S tier — SWE-bench Verified 60–70% ──
   ['openai/gpt-oss-120b', 'GPT OSS 120B', 'S', '62.4%', '128k'],
   ['meta/llama-4-maverick-17b-128e-instruct', 'Llama 4 Maverick', 'S+', '74.8%', '1M'],
-  ['mistralai/mistral-medium-3.5-128b', 'Mistral Medium 3.5', 'S+', '77.6%', '128k'],
+  ['mistralai/mistral-medium-3.5-128b', 'Mistral Medium 3.5', 'S+', '77.6%', '256k'],
   ['mistralai/mistral-small-4-119b-2603', 'Mistral Small 4', 'S', '60.0%', '256k'],
   ['minimaxai/minimax-m3', 'MiniMax M3', 'S+', '78.4%', '1M'],
-  ['qwen/qwen3-coder-480b-a35b-instruct', 'Qwen3 Coder 480B', 'S', '69.6%', '262k'],
-  ['nvidia/mistral-nemotron', 'Mistral Nemotron', 'S', '-', '-'],
+  ['nvidia/mistral-nemotron', 'Mistral Nemotron', 'S', '-', '128k'],
   ['deepseek-ai/deepseek-v3.2', 'DeepSeek V3.2', 'S+', '70.0%', '160k'],
   // ── A+ tier — SWE-bench Verified 50–60% ──
   ['mistralai/mistral-large-3-675b-instruct-2512', 'Mistral Large 675B', 'A+', '58.0%', '256k'],
   ['nvidia/nemotron-3-super-120b-a12b', 'Nemotron 3 Super', 'S', '60.5%', '128k'],
-  ['nvidia/nemotron-3-nano-omni-30b-a3b-reasoning', 'Nemotron 3 Omni', 'A+', '52.0%', '128k'],
+  ['nvidia/nemotron-3-nano-omni-30b-a3b-reasoning', 'Nemotron 3 Omni', 'A+', '52.0%', '256k'],
   ['meta-llama/llama-4-scout-17b-16e-instruct', 'Llama 4 Scout', 'B', '28.0%', '10M'],
   ['nvidia/llama-3.3-nemotron-super-49b-v1.5', 'Llama 3.3 Nemotron Super 49B v1.5', 'A+', '-', '128k'],
   // ── A tier — SWE-bench Verified 40–50% ──
@@ -68,7 +67,7 @@ export const nvidiaNim = [
   ['openai/gpt-oss-20b', 'GPT OSS 20B', 'A+', '50.3%', '128k'],
   ['google/gemma-4-31b-it', 'Gemma 4 31B', 'A+', '52.0%', '256k'],
   ['qwen/qwen2.5-coder-32b-instruct', 'Qwen2.5 Coder 32B', 'A', '47.0%', '128k'],
-  ['mistralai/magistral-small-2506', 'Magistral Small 2506', 'A', '45.0%', '128k'],
+  ['deepseek-ai/deepseek-r1', 'DeepSeek R1', 'A', '49.2%', '128k'],
   ['nvidia/nemotron-3-nano', 'Nemotron 3 Nano', 'A-', '38.8%', '256k'],
   // ── A- tier — SWE-bench Verified 35–40% ──
   ['bytedance/seed-oss-36b-instruct', 'Seed OSS 36B', 'A+', '56.0%', '32k'],
@@ -97,7 +96,8 @@ export const groq = [
 // 📖 Free API keys available at https://cloud.cerebras.ai
 export const cerebras = [
   // ── S+ tier — SWE-bench Verified ≥70% ──
-  ['zai-glm-4.7', 'GLM 4.7', 'S+', '73.8%', '128k'],
+  // ⚠️ DEPRECATED — shutdown 2026-08-17 (Cerebras official notice)
+  ['zai-glm-4.7', 'GLM 4.7', 'S+', '73.8%', '128k', null, '2026-08-17'],
   // ── S tier — SWE-bench Verified 60–70% ──
   ['gpt-oss-120b', 'GPT OSS 120B', 'S', '62.4%', '128k'],
   // ── A tier — SWE-bench Verified 40–50% ──
@@ -138,7 +138,6 @@ export const openrouter = [
   ['poolside/laguna-xs.2:free', 'Poolside Laguna XS.2', 'S', '68.2%', '262k'],
   ['poolside/laguna-xs-2.1:free', 'Poolside Laguna XS 2.1', 'S+', '70.9%', '262k'],
   // ── S tier — SWE-bench Verified 60–70% ──
-  ['openai/gpt-oss-120b:free', 'GPT OSS 120B', 'S', '62.4%', '131k'],
   ['cohere/north-mini-code:free', 'North Mini Code', 'S', '-', '256k'],
   ['tencent/hy3:free', 'Tencent Hy3', 'S', '-', '262k'],
   ['qwen/qwen3-coder:free', 'Qwen3 Coder', 'S', '69.6%', '1M'],
@@ -146,7 +145,6 @@ export const openrouter = [
   ['nvidia/nemotron-3-super-120b-a12b:free', 'Nemotron 3 Super', 'S', '60.5%', '1M'],
   ['nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free', 'Nemotron 3 Omni', 'A+', '52.0%', '256k'],
   // ── A tier — SWE-bench Verified 40–50% ──
-  ['openai/gpt-oss-20b:free', 'GPT OSS 20B', 'A+', '50.3%', '131k'],
   ['nvidia/nemotron-3-nano-30b-a3b:free', 'Nemotron Nano 30B', 'A-', '38.8%', '256k'],
   ['nvidia/nemotron-nano-12b-v2-vl:free', 'Nemotron Nano 12B VL', 'A', '20.0%', '128k'],
   ['google/gemma-4-31b-it:free', 'Gemma 4 31B', 'A+', '52.0%', '262k'],
@@ -235,8 +233,8 @@ export const mistral = [
 // 📖 Free coding model — 30 req/min, 2000/day (phone number required for key)
 // 📖 API keys now use the Mistral platform key format; CODESTRAL_API_KEY remains supported as an alias.
 export const codestral = [
-  // ── B+ tier — SWE-bench Verified 30–35% ──
-  ['codestral-2508', 'Codestral', 'A', '40.0%', '128k'],
+  // ── A tier — SWE-bench Verified 40–50% ──
+  ['codestral-2508', 'Codestral', 'A', '40.0%', '256k'],
   ['codestral-2', 'Codestral 2', 'B+', '-', '128k'],
 ]
 
@@ -387,8 +385,9 @@ export const opencodeZen = [
   ['big-pickle',                       'Big Pickle',              'S+', '72.0%', '200k'],
   ['deepseek-v4-flash-free',           'DeepSeek V4 Flash Free',  'S+', '79.0%', '200k'],
   ['mimo-v2.5-free',                   'MiMo-V2.5 Free',          'S+', '-',     '200k'],
-  ['nemotron-3-ultra-free',            'Nemotron 3 Ultra Free',   'S+', '71.9%',     '200k'],
+  ['nemotron-3-ultra-free',            'Nemotron 3 Ultra Free',   'S+', '71.9%', '200k'],
   ['north-mini-code-free',             'North Mini Code Free',    'B+', '-',     '200k'],
+  ['hy3-free',                         'Tencent Hy3 Free',        'S',  '-',     '200k'],
 ]
 
 // 📖 Kilo source - https://api.kilo.ai/api/gateway
@@ -403,8 +402,10 @@ export const kilo = [
 // 📖 Pro-tagged models from /v1/models are intentionally excluded.
 export const llm7 = [
   // 📖 LLM7 live /v1/models: only `turbo` tier is free (noKeyNeeded). All `pro` models are usage-based paid.
-  // ── S tier — SWE-bench Verified 60–70% ──
-  ['devstral-small-2:24b', 'Devstral Small 2', 'S', '68.0%', '255k'],
+  // ── S+ tier — SWE-bench Verified ≥70% ──
+  ['minimax-m2.7', 'MiniMax M2.7', 'S+', '78.0%', '180k'],
+  // ── A+ tier — SWE-bench Verified 50–60% ──
+  ['gpt-oss:20b', 'GPT OSS 20B', 'A+', '50.3%', '128k'],
   // ── A tier — SWE-bench Verified 40–50% ──
   ['codestral-latest', 'Codestral Latest', 'A', '40.0%', '32k'],
 ]
@@ -440,49 +441,34 @@ export const routeway = [
 // 📖 Test/dev/placeholder zero-price IDs were intentionally excluded.
 export const novita = [
   // ── S+ tier — SWE-bench Verified ≥70% ──
-  ['qwen/qwen3.6-plus', 'Qwen3.6 Plus', 'S+', '78.8%', '1M'],
-  // ── S tier — SWE-bench Verified 60–70% ──
   ['qwen/qwen3.5-plus', 'Qwen3.5 Plus', 'S+', '80.0%', '1M'],
-  ['nex-agi/nex-n2-pro', 'Nex N2 Pro', 'S', '-', '262k'],
-  ['minimax/m2-her', 'MiniMax M2 HER', 'S', '-', '66k'],
+  // ── S tier — SWE-bench Verified 60–70% ──
+  ['tencent/hy3', 'Tencent Hy3', 'S', '-', '262k'],
 ]
 
-// 📖 Ollama Cloud source - https://ollama.com/pricing and https://ollama.com/v1/models
+// 📖 Ollama Cloud source - https://ollama.com/pricing and https://ollama.com/search?c=cloud
 // 📖 Free plan includes cloud model access with session/weekly limits. This list keeps coding-relevant cloud models only.
+// 📖 Catalog verified 2026-07-18 against official Ollama cloud model search page.
 export const ollamaCloud = [
   // ── S+ tier — SWE-bench Verified ≥70% ──
-  ['devstral-2:123b', 'Devstral 2 123B', 'S+', '72.2%', '256k'],
-  ['qwen3-coder:480b', 'Qwen3 Coder 480B', 'S', '69.6%', '256k'],
   ['nemotron-3-ultra', 'Nemotron 3 Ultra', 'S+', '71.9%', '256k'],
-  ['glm-4.7', 'GLM 4.7', 'S+', '73.8%', '200k'],
+  ['glm-5.1', 'GLM 5.1', 'S+', '82.8%', '128k'],
+  ['glm-5.2', 'GLM 5.2', 'S+', '82.8%', '128k'],
+  ['minimax-m2.7', 'MiniMax M2.7', 'S+', '78.0%', '200k'],
+  ['minimax-m3', 'MiniMax M3', 'S+', '78.4%', '512k'],
+  ['minimax-m2.5', 'MiniMax M2.5', 'S+', '80.2%', '200k'],
+  ['kimi-k2.6', 'Kimi K2.6', 'S+', '80.2%', '262k'],
+  ['deepseek-v4-flash', 'DeepSeek V4 Flash', 'S+', '79.0%', '256k'],
+  ['deepseek-v4-pro', 'DeepSeek V4 Pro', 'S+', '80.6%', '256k'],
   // ── S tier — SWE-bench Verified 60–70% ──
-  ['qwen3-coder-next', 'Qwen3 Coder Next', 'S+', '70.6%', '256k'],
+  ['kimi-k2.7-code', 'Kimi K2.7 Code', 'S', '60.4%', '262k'],
   ['gpt-oss:120b', 'GPT OSS 120B', 'S', '62.4%', '128k'],
-  ['minimax-m3', 'MiniMax M3', 'S+', '78.4%', '512k'],
-  // ── A+ tier — SWE-bench Verified 50–60% ──
   ['nemotron-3-super', 'Nemotron 3 Super', 'S', '60.5%', '256k'],
-  ['cogito-2.1:671b', 'Cogito 2.1 671B', 'A+', '-', '160k'],
-  // ── A tier — SWE-bench Verified 40–50% ──
+  // ── A+ tier — SWE-bench Verified 50–60% ──
+  ['kimi-k2.5', 'Kimi K2.5', 'A+', '-', '200k'],
   ['gemma4:31b', 'Gemma 4 31B', 'A+', '52.0%', '256k'],
   ['gpt-oss:20b', 'GPT OSS 20B', 'A+', '50.3%', '128k'],
-  ['glm-4.6', 'GLM 4.6', 'A', '-', '200k'],
-  ['qwen3-next:80b', 'Qwen3 Next 80B', 'S+', '70.6%', '256k'],
-  ['qwen3-vl:235b', 'Qwen3 VL 235B', 'A', '-', '256k'],
-  ['qwen3-vl:235b-instruct', 'Qwen3 VL 235B Instruct', 'A', '-', '256k'],
-  // ── A- tier — SWE-bench Verified 35–40% ──
-  ['devstral-small-2:24b', 'Devstral Small 2 24B', 'S', '68.0%', '256k'],
-  ['gemma3:27b', 'Gemma 3 27B', 'B', '22.0%', '128k'],
-  ['minimax-m2', 'MiniMax M2', 'S', '69.4%', '200k'],
-  ['minimax-m2.1', 'MiniMax M2.1', 'S+', '74.0%', '200k'],
-  ['minimax-m2.5', 'MiniMax M2.5', 'S+', '80.2%', '200k'],
-  ['ministral-3:8b', 'Ministral 3 8B', 'A-', '-', '256k'],
-  ['ministral-3:14b', 'Ministral 3 14B', 'A-', '-', '256k'],
   ['nemotron-3-nano:30b', 'Nemotron 3 Nano 30B', 'A-', '38.8%', '1M'],
-  // ── B+ tier — SWE-bench Verified 30–35% ──
-  ['gemma3:4b', 'Gemma 3 4B', 'B+', '-', '128k'],
-  ['gemma3:12b', 'Gemma 3 12B', 'B+', '-', '128k'],
-  ['ministral-3:3b', 'Ministral 3 3B', 'B+', '-', '256k'],
-  ['rnj-1:8b', 'RNJ 1 8B', 'B+', '-', '32k'],
 ]
 
 // 📖 All sources combined - used by the main script
@@ -598,11 +584,14 @@ export const sources = {
 
 // 📖 Flatten all models from all sources — each entry includes providerKey as 6th element
 // 📖 providerKey lets the main CLI know which API key and URL to use per model
+// 📖 Models with a deprecatedAfter date (7th tuple element) are auto-filtered after that date
 export const MODELS = [];
+const _today = new Date().toISOString().split('T')[0];
 for (const [sourceKey, sourceData] of Object.entries(sources)) {
   if (!sourceData || !sourceData.models) continue
   for (const model of sourceData.models) {
-    const [modelId, label, tier, sweScore, ctx, addedDate] = model
+    const [modelId, label, tier, sweScore, ctx, addedDate, deprecatedAfter] = model
+    if (deprecatedAfter && _today > deprecatedAfter) continue
     MODELS.push([modelId, label, tier, sweScore, ctx, sourceKey, addedDate || null])
   }
 }

package/src/core/ping.js

@@ -205,6 +205,9 @@ export function extractQuotaPercent(headers) {
   const variants = [
     ['x-ratelimit-remaining', 'x-ratelimit-limit'],
     ['x-ratelimit-remaining-requests', 'x-ratelimit-limit-requests'],
+    ['x-ratelimit-remaining-requests-day', 'x-ratelimit-limit-requests-day'],
+    ['x-ratelimit-remaining-tokens', 'x-ratelimit-limit-tokens'],
+    ['x-ratelimit-remaining-tokens-minute', 'x-ratelimit-limit-tokens-minute'],
     ['ratelimit-remaining', 'ratelimit-limit'],
     ['ratelimit-remaining-requests', 'ratelimit-limit-requests'],
   ]

package/src/core/quota-capabilities.js

@@ -34,18 +34,20 @@
 
 /** @type {Record<string, ProviderCapability>} */
 export const PROVIDER_CAPABILITIES = {
-  // Providers that return x-ratelimit-remaining / x-ratelimit-limit headers
-  nvidia: { telemetryType: 'header', supportsEndpoint: false, usageDisplay: 'ok',      resetCadence: 'none' },
+  // Providers with verified quota telemetry from official docs or live API behavior.
+  nvidia: { telemetryType: 'unknown', supportsEndpoint: false, usageDisplay: 'ok', resetCadence: 'none' },
   groq: { telemetryType: 'header', supportsEndpoint: false, usageDisplay: 'percent', resetCadence: 'daily' },
-  cerebras: { telemetryType: 'header', supportsEndpoint: false, usageDisplay: 'percent', resetCadence: 'unknown' },
-  sambanova: { telemetryType: 'header', supportsEndpoint: false, usageDisplay: 'percent', resetCadence: 'unknown' },
-  'github-models': { telemetryType: 'header', supportsEndpoint: false, usageDisplay: 'percent', resetCadence: 'unknown' },
-  mistral: { telemetryType: 'header', supportsEndpoint: false, usageDisplay: 'percent', resetCadence: 'monthly' },
-  scaleway: { telemetryType: 'header', supportsEndpoint: false, usageDisplay: 'percent', resetCadence: 'unknown' },
-  googleai: { telemetryType: 'header', supportsEndpoint: false, usageDisplay: 'percent', resetCadence: 'daily' },
-  codestral: { telemetryType: 'header', supportsEndpoint: false, usageDisplay: 'percent', resetCadence: 'daily' },
-  qwen: { telemetryType: 'header', supportsEndpoint: false, usageDisplay: 'percent', resetCadence: 'unknown' },
-  ovhcloud: { telemetryType: 'header', supportsEndpoint: false, usageDisplay: 'percent', resetCadence: 'unknown' },
+  cerebras: { telemetryType: 'header', supportsEndpoint: false, usageDisplay: 'percent', resetCadence: 'daily' },
+
+  // Providers not yet proven to expose reliable remaining-quota telemetry.
+  sambanova: { telemetryType: 'unknown', supportsEndpoint: false, usageDisplay: 'ok', resetCadence: 'unknown' },
+  'github-models': { telemetryType: 'unknown', supportsEndpoint: false, usageDisplay: 'ok', resetCadence: 'unknown' },
+  mistral: { telemetryType: 'unknown', supportsEndpoint: false, usageDisplay: 'ok', resetCadence: 'monthly' },
+  scaleway: { telemetryType: 'unknown', supportsEndpoint: false, usageDisplay: 'ok', resetCadence: 'unknown' },
+  googleai: { telemetryType: 'unknown', supportsEndpoint: false, usageDisplay: 'ok', resetCadence: 'daily' },
+  codestral: { telemetryType: 'unknown', supportsEndpoint: false, usageDisplay: 'ok', resetCadence: 'daily' },
+  qwen: { telemetryType: 'unknown', supportsEndpoint: false, usageDisplay: 'ok', resetCadence: 'unknown' },
+  ovhcloud: { telemetryType: 'unknown', supportsEndpoint: false, usageDisplay: 'ok', resetCadence: 'unknown' },
 
   // Providers that have a dedicated usage/credits endpoint
   openrouter: { telemetryType: 'endpoint', supportsEndpoint: true, usageDisplay: 'percent', resetCadence: 'unknown' },
@@ -56,7 +58,7 @@ export const PROVIDER_CAPABILITIES = {
   'opencode-zen': { telemetryType: 'unknown', supportsEndpoint: false, usageDisplay: 'ok', resetCadence: 'unknown' },
   kilo: { telemetryType: 'unknown', supportsEndpoint: false, usageDisplay: 'ok', resetCadence: 'unknown' },
   llm7: { telemetryType: 'unknown', supportsEndpoint: false, usageDisplay: 'ok', resetCadence: 'rolling' },
-  routeway: { telemetryType: 'header', supportsEndpoint: false, usageDisplay: 'percent', resetCadence: 'unknown' },
+  routeway: { telemetryType: 'unknown', supportsEndpoint: false, usageDisplay: 'ok', resetCadence: 'unknown' },
   novita: { telemetryType: 'unknown', supportsEndpoint: false, usageDisplay: 'ok', resetCadence: 'unknown' },
   'ollama-cloud': { telemetryType: 'unknown', supportsEndpoint: false, usageDisplay: 'ok', resetCadence: 'rolling' },
 }

package/src/core/router-daemon.js

@@ -31,7 +31,7 @@
 
 import { createServer } from 'node:http'
 import { existsSync, readFileSync } from 'node:fs'
-import { dirname, join, resolve as resolvePath } from 'node:path'
+import { dirname, join, resolve as resolvePath, sep as pathSep } from 'node:path'
 import { fork } from 'node:child_process'
 import { randomUUID } from 'node:crypto'
 import { appendFileSync, renameSync, statSync, unlinkSync, writeFileSync } from 'node:fs'
@@ -434,7 +434,7 @@ function serveWebStaticFile(res, pathname, requestId) {
   // 📖 Without this, `pathname` like `/../../etc/passwd` escapes the dist root.
   const requested = pathname === '/' ? 'index.html' : pathname.replace(/^\/+/, '')
   const candidate = resolvePath(WEB_DIST_DIR, requested)
-  if (candidate !== WEB_DIST_DIR && !candidate.startsWith(WEB_DIST_DIR + '/')) {
+  if (candidate !== WEB_DIST_DIR && !candidate.startsWith(WEB_DIST_DIR + pathSep)) {
     sendError(res, 403, 'Forbidden', 'invalid_request_error', 'path_traversal_blocked', requestId)
     return
   }
@@ -454,7 +454,7 @@ function serveWebStaticFile(res, pathname, requestId) {
 
   if (stats.isDirectory()) {
     const dirIndex = resolvePath(candidate, 'index.html')
-    if (dirIndex.startsWith(WEB_DIST_DIR + '/') && existsSync(dirIndex)) {
+    if (dirIndex.startsWith(WEB_DIST_DIR + pathSep) && existsSync(dirIndex)) {
       serveStaticFromDist(res, dirIndex)
       return
     }

package/src/tui/app.js

@@ -313,6 +313,7 @@ export async function runApp(cliArgs, config, startupOptions = {}) {
       status: 'pending',
       pings: [],  // 📖 All ping results (ms or 'TIMEOUT')
       httpCode: null,
+      hasApiKey: !!getApiKey(config, providerKey),
       isPinging: false, // 📖 Per-row live flag so Last Ping can keep last value and show a spinner during refresh.
       hidden: false,  // 📖 Simple flag to hide/show models
     }))
@@ -431,6 +432,7 @@ export async function runApp(cliArgs, config, startupOptions = {}) {
 
     try {
       const providerApiKey = getApiKey(state.config, r.providerKey) ?? null
+      r.hasApiKey = !!providerApiKey
       const providerUrl = sources[r.providerKey]?.url ?? sources.nvidia.url
       let { code, ms, quotaPercent } = await ping(providerApiKey, r.modelId, r.providerKey, providerUrl)
 

package/src/tui/render-table.js

@@ -28,7 +28,7 @@
  *   - ../src/constants.js: PING_INTERVAL, FRAMES
  *   - ../src/tier-colors.js: TIER_COLOR
  *   - ../src/utils.js: getAvg, getVerdict, getUptime, getStabilityScore
- *   - ../src/ping.js: usagePlaceholderForProvider
+ *   - ../src/quota-capabilities.js: supportsUsagePercent
  *   - ../src/render-helpers.js: calculateViewport, sortResultsWithPinnedFavorites, padEndDisplay, fadedRow
  *
  *   @see bin/free-coding-models.js — main entry point that calls renderTable
@@ -50,7 +50,7 @@ import {
 import { themeColors, currentPalette, getProviderRgb, getTierRgb, getReadableTextRgb, getTheme, THEME_BG_RGB } from './theme.js'
 import { TIER_COLOR } from './tier-colors.js'
 import { getAvg, getVerdict, getUptime, getStabilityScore, getVersionStatusInfo, isNewModel } from '../core/utils.js'
-import { usagePlaceholderForProvider } from '../core/ping.js'
+import { supportsUsagePercent } from '../core/quota-capabilities.js'
 import { formatBenchmarkLatency, formatBenchmarkTps } from '../core/benchmark.js'
 import { calculateViewport, sortResultsWithPinnedFavorites, padEndDisplay, displayWidth, stripAnsi, fadedRow } from './render-helpers.js'
 import { getToolMeta, TOOL_METADATA, TOOL_MODE_ORDER, isModelCompatibleWithTool } from '../core/tool-metadata.js'
@@ -96,6 +96,7 @@ const COLUMN_SORT_MAP = {
   uptime: 'uptime',
   aiLatency: 'aiLatency',
   tps: 'tps',
+  quota: 'usage',
 }
 export { COLUMN_SORT_MAP }
 
@@ -284,7 +285,7 @@ export function renderTable({
   const W_RANK = 6
   const W_TIER = 5
   const W_CTX = 4
-  const W_SOURCE = 14
+  const W_SOURCE = 11
   const W_MODEL = 26
   const W_SWE = 5
   const W_STATUS = 17
@@ -292,9 +293,9 @@ export function renderTable({
   const W_UPTIME = 6
   const W_AI_LATENCY = 17
   const W_TPS = 5
+  const W_QUOTA = 6
 
   // const W_TOKENS = 7 // Used column removed
-  // const W_USAGE = 7 // Usage column removed
   const MIN_TABLE_WIDTH = WIDTH_WARNING_MIN_COLS
 
   // 📖 Responsive column visibility: progressively hide least-useful columns
@@ -308,6 +309,7 @@ export function renderTable({
   let wAvg = 9
   let wStab = 11
   let wSource = W_SOURCE
+  let wModel = W_MODEL
   let wStatus = W_STATUS
   let wAiLatency = W_AI_LATENCY
   let showRank = true
@@ -324,10 +326,11 @@ export function renderTable({
       cols.push(W_MOOD)
       if (showRank) cols.push(W_RANK)
       if (showTier) cols.push(W_TIER)
-      cols.push(W_SWE, W_CTX, W_MODEL, wSource, wPing, wAvg, wStatus, W_VERDICT)
+      cols.push(W_SWE, W_CTX, wModel, wSource, wPing, wAvg, wStatus, W_VERDICT)
       if (showStability) cols.push(wStab)
       if (showUptime) cols.push(W_UPTIME)
       if (showBenchmarkColumns) cols.push(wAiLatency, W_TPS)
+      cols.push(W_QUOTA)
       return ROW_MARGIN + cols.reduce((a, b) => a + b, 0) + (cols.length - 1) * SEP_W
     }
 
@@ -338,8 +341,9 @@ export function renderTable({
       wAvg = 9       // 'Avg Ping' stays aligned with Last Ping
       wStab = 8      // 'StaB.' instead of 'Stability'
       wSource = 7    // Provider truncated to 4 chars + '…', 7 cols total
+      wModel = 21    // Keep compact tables wide enough for Rank + Quota
       wStatus = 13   // Health truncated after 6 chars + '…'
-      wAiLatency = 13 // Mirror compact Health text when health is not good
+      wAiLatency = 9 // Keep compact tables wide enough for Rank + Quota
     }
     // 📖 Steps 2–6: Progressive column hiding (least useful first)
     if (calcWidth() > terminalCols) showRank = false
@@ -359,7 +363,7 @@ export function renderTable({
     if (showTier) colDefs.push({ name: 'tier', width: W_TIER })
     colDefs.push({ name: 'swe', width: W_SWE })
     colDefs.push({ name: 'ctx', width: W_CTX })
-    colDefs.push({ name: 'model', width: W_MODEL })
+    colDefs.push({ name: 'model', width: wModel })
     colDefs.push({ name: 'source', width: wSource })
     colDefs.push({ name: 'ping', width: wPing })
     colDefs.push({ name: 'avg', width: wAvg })
@@ -371,6 +375,7 @@ export function renderTable({
       colDefs.push({ name: 'aiLatency', width: wAiLatency })
       colDefs.push({ name: 'tps', width: W_TPS })
     }
+    colDefs.push({ name: 'quota', width: W_QUOTA })
     let x = ROW_MARGIN + 1 // 📖 1-based: first column starts after the 2-char left margin
     const columns = []
     for (let i = 0; i < colDefs.length; i++) {
@@ -517,7 +522,7 @@ export function renderTable({
     if (headerFlashColumn === 'tier') return flashHeader(tierLabel, W_TIER)
     return colorFirst(tierLabel, W_TIER)
   })()
-  const modelH_c   = headerStyle('model', modelLabel, W_MODEL)
+  const modelH_c   = headerStyle('model', modelLabel, wModel)
   const sweH_c     = headerStyle('swe', sweLabel, W_SWE)
   const ctxH_c     = headerStyle('ctx', ctxLabel, W_CTX)
   const pingH_c    = headerStyle('ping', pingLabel, wPing)
@@ -580,8 +585,15 @@ export function renderTable({
     const padding = ' '.repeat(Math.max(0, W_TPS - plain.length))
     return themeColors.dim(plain + padding)
   })()
+  const quotaH_c = (() => {
+    if (headerFlashColumn === 'quota' || headerFlashColumn === 'usage') {
+      const ft = (dir + ' Quota').length <= W_QUOTA ? dir + ' Quota' : 'Quota' + dir
+      return flashHeader(ft, W_QUOTA)
+    }
+    if (sortColumn === 'usage') return sortActiveHeader('Quota', W_QUOTA)
+    return themeColors.dim('Quota'.padEnd(W_QUOTA))
+  })()
 
-  // 📖 Usage column removed from UI – no header or separator for it.
   // 📖 Header row: conditionally include columns based on responsive visibility
   const headerParts = [moodH_c]
   if (showRank) headerParts.push(rankH_c)
@@ -590,6 +602,7 @@ export function renderTable({
   if (showStability) headerParts.push(stabH_c)
   if (showUptime) headerParts.push(uptimeH_c)
   if (showBenchmarkColumns) headerParts.push(aiLatencyH_c, tpsH_c)
+  headerParts.push(quotaH_c)
   lines.push('  ' + headerParts.join(COL_SEP))
 
   // 📖 Mouse support: the column header row is the last line we just pushed.
@@ -666,7 +679,7 @@ export function renderTable({
       favoritePrefix = '🆕 '
     }
     const prefixDisplayWidth = displayWidth(favoritePrefix)
-    const nameWidth = Math.max(0, W_MODEL - prefixDisplayWidth)
+    const nameWidth = Math.max(0, wModel - prefixDisplayWidth)
     const name = favoritePrefix + r.label.slice(0, nameWidth).padEnd(nameWidth)
     const sweScore = r.sweScore ?? '—'
     // 📖 SWE% colorized on the same gradient as Tier:
@@ -901,9 +914,23 @@ export function renderTable({
     // 📖 Check if this model is incompatible with the active tool mode
     const isIncompatible = !isModelCompatibleWithTool(r.providerKey, mode)
 
-    // 📖 Usage column removed from UI – no usage data displayed.
-    // (We keep the logic but do not render it.)
-    const usageCell = ''
+    // 📖 Quota column: provider telemetry if available, N/A when impossible,
+    // 📖 and '-' when the provider is not configured.
+    let quotaCell
+    if (!r.hasApiKey) {
+      quotaCell = themeColors.dim('-'.padEnd(W_QUOTA))
+    } else if (!supportsUsagePercent(r.providerKey)) {
+      quotaCell = themeColors.dim('N/A'.padEnd(W_QUOTA))
+    } else if (typeof r.usagePercent === 'number' && Number.isFinite(r.usagePercent)) {
+      const quotaText = `${Math.round(r.usagePercent)}%`.padEnd(W_QUOTA)
+      quotaCell = r.usagePercent >= 50
+        ? themeColors.metricGood(quotaText)
+        : r.usagePercent >= 20
+          ? themeColors.metricWarn(quotaText)
+          : themeColors.metricBad(quotaText)
+    } else {
+      quotaCell = themeColors.dim('…'.padEnd(W_QUOTA))
+    }
 
     // 📖 AI Latency + TPS columns — same benchmark result, split into two readable metrics.
     // 📖 Benchmark results are shown regardless of health status (up/timeout/down/429/noauth).
@@ -960,6 +987,7 @@ export function renderTable({
     if (showStability) rowParts.push(stabCell)
     if (showUptime) rowParts.push(uptimeCell)
     if (showBenchmarkColumns) rowParts.push(latencyCell, tpsCell)
+    rowParts.push(quotaCell)
     const row = '  ' + rowParts.join(COL_SEP)
 
     // 📖 "Unusable" models (no key / bad key) are visually de-emphasized at the