Tencent’s Hy3 tops OpenRouter with 8.98T tokens and a solid coding test

alex_verem · x · 2026-07-23

Tencent’s Hy3 is now the most-used model on OpenRouter this week, with 8.98T tokens and a sharp lead over GPT, Claude and DeepSeek.

• The author tested it with a coding prompt asking for a brute-force solution, an O(n) optimization, and a complexity explanation.
• Hy3 correctly produced the O(n²) approach first, then rewrote it into a single-pass O(n) version tracking the minimum price.
• The post highlights published specs: 295B total parameters, 21B active per token, 256K context, mixed fast/slow reasoning, Apache 2.0 licensing, and pricing listed by Tencent Cloud at 1 CNY/M input tokens and 4 CNY/M output tokens.
• The author notes that a single coding example is not a benchmark, but the usage surge may reflect those specs.

Original post →

More from Models

• Local benchmark: Flint matches Qwen3.6 quality at 50/52 while running 3.6x faster on M4 Pro — RichardSocher · 2026-10-27
• ICML 2026 position paper: LLM 'thinking traces' are a harmful anthropomorphism — rao2z · 2026-10-27
• One stray 'u' and Claude starts spelling like a Brit — cantrell · 2026-10-27
• Cohere's Invent-A-Dataset paper: zero-seed synthetic data beats frontier APIs on quality and diversity — idanbeck · 2026-10-27
• ARC Prize redesigns leaderboard: highest reasoning effort by default, CSV export — mhmazur · 2026-10-27
• Cohere: open-weight Qwen3.8-Flash-Next tried uploading probe file to Beijing cloud bucket in sandbox — wstone_bd · 2026-10-27

Models channel →