Grok 4 Fast brings frontier search reasoning to free users
SpaceXAI released Grok 4 Fast to all Grok users and through low-cost APIs. The unified reasoning model combines a two-million-token context window with native web, X, and code-tool use, while its model card reports agent-abuse, prompt-injection, deception, and dual-use evaluations.
0 comments · 0 votes
Sign in to join the discussion →
No comments yet. Start the discussion.
Why it moved the index
Free consumer access and inexpensive API access materially diffuse a near-frontier reasoning and search model with native tool use. The primary launch and model card establish access and safeguards, but most capability comparisons are developer-reported benchmarks, so confidence is below independently replicated evidence.
Assessment history
-
R1
Toward 58 · confidence 87
New dated primary-source release and model-card evidence absent from the durable catalogue.
14 Aug 2026
Share this page
-
DoomBench assesses “Grok 4 Fast brings frontier search reasoning to free users” as evidence moving toward doom, with magnitude 58 and confidence 87 out of 100 in the deployment reach category.
-
The DoomBench assessment of “Grok 4 Fast brings frontier search reasoning to free users” is based on reporting from SpaceXAI and records the editorial rationale, source quality, attribution, and revision history.
-
DoomBench summarizes “Grok 4 Fast brings frontier search reasoning to free users” as follows: SpaceXAI released Grok 4 Fast to all Grok users and through low-cost APIs. The unified reasoning model combines a two-million-token context...
https://www.doombench.com/news/grok-4-fast-brings-frontier-search-reasoning-to-free-users-2025-09-19