Image Credit: xAI News Website
xAI releases Grok 4 Fast — frontier AI at 98% lower cost. Benchmarks, context, and search dominance make it a game-changer for enterprises and users.
xAI unveiled Grok 4 Fast, a transformative advancement in artificial intelligence that prioritizes cost-efficient reasoning without compromising frontier-level performance. Built upon the foundation of Grok 4, this model delivers exceptional capabilities across enterprise and consumer applications, optimized for token efficiency. By blending state-of-the-art intelligence density with accessibility, Grok 4 Fast redefines the boundaries of affordable AI, enabling businesses, developers, and individuals to harness advanced reasoning at a fraction of the cost.
Unified Architecture for Seamless Performance
A hallmark of Grok 4 Fast is its unified architecture, which integrates reasoning (extended chain-of-thought processes) and non-reasoning (rapid responses) within a single set of model weights. Unlike predecessors requiring distinct models, this approach uses system prompts to dynamically adjust behavior, reducing end-to-end latency and token consumption. For users on grok.com, this translates to fluid interactions—swift answers for straightforward queries and rigorous analysis for complex challenges. Developers leveraging the xAI API can fine-tune this flexibility, optimizing for speed or depth based on specific use cases. The model is offered in two variants: grok-4-fast-reasoning and grok-4-fast-non-reasoning, both supporting a 2 million token context window.
Benchmark Performance: Efficiency Meets Excellence
Grok 4 Fast sets a new standard in reasoning benchmarks, achieving near-parity with Grok 4 while using 40% fewer thinking tokens on average. This efficiency yields a 98% reduction in cost to match Grok 4’s performance on frontier tasks, as validated by Artificial Analysis. The model’s state-of-the-art price-to-intelligence ratio outshines competitors like GPT-5, Gemini 2.5 Flash, and Claude 4 on the Artificial Analysis Intelligence Index. Below are key pass@1 benchmark results (without tools):
|
Benchmark |
Grok 4 Fast |
Grok 4 |
Grok 3 Mini (High) |
GPT-5 (High) |
GPT-5 Mini (High) |
|---|---|---|---|---|---|
|
GPQA Diamond |
85.7% | 87.5% | 79.0% | 85.7% | 82.3% |
|
AIME 2025 |
92.0% | 91.7% | 83.0% | 94.6% | 91.1% |
|
HMMT 2025 |
93.3% | 90.0% | 74.0% | 93.3% | 87.8% |
|
HLE |
20.0% | 25.4% | 11.0% | 24.8% | 16.7% |
|
LiveCodeBench (Jan-May) |
80.0% | 79.0% | 70.0% | 86.8% | 77.4% |
These results showcase Grok 4 Fast’s prowess in graduate-level physics (GPQA Diamond), competitive mathematics (AIME and HMMT), high-level evaluations (HLE), and coding challenges (LiveCodeBench). Notably, it achieves 92.0% on AIME 2025, surpassing Grok 4, with only 28,000 thinking tokens, as illustrated in xAI’s performance charts.
Intelligence Density: Maximum Value, Minimum Cost
The model’s intelligence density—maximum performance at minimal cost—is a game-changer. On the Artificial Analysis Intelligence Index, Grok 4 Fast scores approximately 75 while operating at costs as low as $16 (on a log scale up to $4,096), making it 47 times cheaper than some rivals. This efficiency stems from large-scale reinforcement learning (RL), optimizing token usage without sacrificing reasoning depth. For enterprises, this translates to scalable AI deployment; for consumers, it means affordable access to cutting-edge intelligence.
Agentic Search and Real-Time Capabilities
Grok 4 Fast excels in agentic tasks, leveraging end-to-end tool-use RL to intuitively invoke code execution or web browsing. Its frontier search capabilities enable seamless navigation of web and X platforms, synthesizing real-time data from links, images, and videos. Benchmark results highlight this strength:
|
Benchmark |
Grok 4 Fast |
Grok 4 |
Grok 3 (No Reasoning) |
|---|---|---|---|
|
BrowseComp |
44.9% |
43.0% | — |
|
SimpleQA |
95.0% |
94.0% |
82.0% |
|
Reka Research Eval |
66.0% |
58.0% |
37.0% |
|
BrowseComp (zh) |
51.2% |
45.0% |
10.8% |
|
X Bench Deepsearch (zh) |
74.0% |
66.0% |
27.0% |
|
X Browse* |
58.0% |
53.2% |
20.8% |
*X Browse is an internal benchmark for multihop search on X.
The model’s 95.0% SimpleQA score and 74.0% on X Bench Deepsearch (Chinese) underscore its multilingual versatility and rapid data synthesis. A practical example illustrates this: when asked, “What is the maximum number of experience points possible in Path of Exile 2?” Grok 4 Fast processes the query in 24 seconds, confirming the max level of 100 requires 4,250,334,444 XP. It methodically searches sources like PoE Wiki and poe2db.tw, verifies cross-game consistency, and calculates cumulative XP thresholds, demonstrating transparent and reliable reasoning.
Dominating General Domain Performance
In LMSYS’s Arena, Grok 4 Fast proves its mettle. Its search variant, grok-4-fast-search (codename: menlo), secures the #1 spot in the Search Arena with a 1163 Elo score, outpacing o3-search by 17 points. In the Text Arena, grok-4-fast (codename: tahoe) ranks #8, matching grok-4-0709 and surpassing all peers in its weight class, where competitors rank 18th or lower. This balance of efficiency and performance makes it ideal for diverse, real-world applications.
Accessibility for All Users
xAI’s commitment to inclusivity shines with Grok 4 Fast, available immediately on grok.com, iOS, and Android apps for all users, including free tiers. In Fast and Auto modes, users experience enhanced search and query handling, providing a faster, high-quality experience. While free access marks a significant step toward democratization, standard platform usage limits may apply depending on the tier.
API Integration and Pricing
For developers, Grok 4 Fast offers grok-4-fast-reasoning and grok-4-fast-non-reasoning via the xAI API, both with a 2M token context window. Pricing is structured for flexibility and detailed in xAI’s official documentation:
|
Token Type |
<128k tokens |
≥128k tokens |
|---|---|---|
|
Input tokens |
$0.20 / 1M |
$0.40 / 1M |
|
Output tokens |
$0.50 / 1M |
$1.00 / 1M |
|
Cached input tokens |
$0.05 / 1M |
— |
This tiered model supports varied applications, from lightweight chatbots to intensive analytics, with cached tokens optimizing long-context tasks.
Future Horizons
xAI plans to refine Grok 4 Fast based on user feedback via x.com, with enhancements in multimodal capabilities and agentic features on the horizon. This iterative approach ensures the model evolves to meet diverse needs, maintaining its edge in the AI landscape.
Grok 4 Fast is more than an upgrade—it’s a paradigm shift in cost-efficient AI reasoning. By delivering frontier performance at a fraction of the cost, xAI empowers enterprises to scale intelligently and individuals to explore confidently. Whether tackling complex benchmarks or answering niche queries, Grok 4 Fast proves that advanced AI can be both powerful and accessible, paving the way for a more inclusive future in artificial intelligence.
Join the Poniak Search early access program.
We’re opening an early access to our AI-Native Poniak Search. The first 500 sign-ups will unlock exclusive future benefits and rewards as we grow.
[Sign up here -> Poniak]
Limited seats available.