Terminal benchmark TB-fn: OpenAI and Anthropic maintain clear lead
abeirami · x · 2026-08-28
Fidian tested top 20 models from Terminal-Bench 2.1 on their TB-fn variant. Results show only OpenAI and Anthropic models remain on the frontier, widening the score gap from 15 to 22 points. GLM-5.3 leads open-weight models but falls further behind Sol max (gap increases from 1.0 to 8.6 points). Six models, including Grok 4.5 and Kimi K3, dropped 6-11 points. Per-attempt costs increased across the board on TB-fn.
More from Models
- Nathan Lambert slams Meta's Muse Open License as 'trash' — natolambert · 2026-10-16
- Google launches Gemini 4 Argon to Ultra and Pro subscribers, touting physics and 3D skills — ShikharMurty · 2026-10-16
- Google gives X users early access to test Argon, not just companies — lyraxana · 2026-10-16
- Grok Voice 3.0 Tops Design Arena Speech Leaderboard, Up 7.4 Points — XFreeze · 2026-10-16
- Robotics LLMs tested on 16 tasks: GPT-6 Astra hits 17.2% success vs Fable's 9.1% — pabbeel · 2026-10-16
- Meta open-sources Muse Spark: top US open model, scores just below DeepSeek — bindureddy · 2026-10-16