HHRI-AI LLM EvalBoard

Comprehensive LLM Benchmark Leaderboard — Traditional Chinese, Multilingual & General Capabilities

A live, interactive leaderboard tracking FoxBrain models against frontier baselines across Traditional Chinese, multilingual, and general-capability benchmarks.

---

🔗 Live leaderboard

👉 https://muhammadsaqlainaslam.github.io/tmmlu-leaderboard/

---

What's inside

• Visual Leaderboard — sortable rankings, discipline radar map, mean accuracy per discipline bar chart with score labels, expandable per-model subject breakdowns
• Side-by-Side Comparison — filterable model comparison table across all benchmark tasks
• Light / Dark / System theme modes — persists across visits
• Google-style search — instant filtering with match highlighting

Benchmarks covered

TMMLU+, TCEval-v2, BigBenchHard, τ-bench, τ²-bench, MRCR v2, AIEC — spanning Traditional Chinese knowledge, reasoning, agentic tool-use, and long-context retrieval.

Models tracked

FoxBrain_v1.2_70B_0630, FoxBrain_v1.5_20251203_remove_repeat_lora, NVIDIA-Nemotron-3-Super-120B-A12B-BF16, NVIDIA-Nemotron-3-Nano-30B-A3B-BF16, gemma-4-31B-it, Gemma-3-TAIDE-12b-Chat-ZH-TW-Prompt, gpt-oss-20b

Related projects

• 📊 FoxBrain EvalHub (frontier benchmark & model reference) — muhammadsaqlainaslam.github.io/foxbrain-eval-dashboard
• 📁 FoxBrain EvalHub GitHub repo — github.com/MuhammadSaqlainAslam/foxbrain-eval-dashboard

---

Curated by Muhammad Saqlain · HHRI-AI / Foxconn AI Research Center

Last updated: June 15, 2026