Summary
According to a post on X by @ValsAI, Terra sits less than four points behind GPT-6.1 Sol on the Vals index, at half Sol's token price. Terra scored 84% on Vibe Code Bench, up nine points on GPT-5.6 Terra, and gained 27 points on Terminal Bench at a third of the cost. On Vals' legal, tax and finance agent benchmarks, Terra often hedged or asked for more information instead of committing to an answer, and used over three times as many output tokens as Sol.
AI-assisted summary based on the listed source.
What happened
Vals puts Terra less than four points behind GPT-6.1 Sol at half Sol's token price, with 84% on Vibe Code Bench (up nine points on GPT-5.6 Terra) and a 27-point jump on Terminal Bench at a third of the cost. On its legal, tax and finance agent benchmarks, though, Terra often hedged or asked for more information...
Why it matters
Developers comparing cheaper models for coding work may find Terra's coding scores and lower prices relevant based on these results. Teams considering AI for legal, tax or finance tasks should note the post's flag on hedging, since it suggests answers may need closer review. Watch for the full Vals breakdown to see whether these gaps hold across more tests.
Signal Intelligence
Signal Strength 90%
Technical label WATCH
Public Interest 50
Category RESEARCH
Reader Depth TECHNICAL
Signal Strength reflects source quality, relevance, freshness and evidence. Public Interest helps organize discovery; it is not proof of truth.
Public Interest components
Recognizable Entity Score 60
Practical Impact Score 8
Novelty Interest Score 94
Consequence Score 46
Curiosity Score 16
Shareability Score 59
Why this is here
VQV surfaced this signal because it is recent, relevant to AI Coding Tools, connected to Scobleizer News Pulse (X).
