whatmodel.sh / Grok 4 Fast
Grok 4 Fast
↗ share
+ contribute
Grok 4 Fast is a closed-weights language model from SpaceXAI (xAI), released September 2025.
Catalogue identity notes
Comparison-only. Arena entries 'grok-4-fast-reasoning' (text, webdev), 'grok-4-fast-chat' (text), 'grok-4-fast-search' (search). Released 2025-09: xAI 'Grok 4 Fast' post dated Sep 19, 2025. xAI: 'Grok 4 Fast introduces a unified architecture where reasoning ... and non-reasoning ... are handled by the same model weights, steered via system prompts', so the reasoning/chat entries are one model. grok-4-fast-reasoning/-non-reasoning retired from the API on May 15, 2026.
$ how it compares
The LLMs ranked either side of Grok 4 Fast, out of 828 with a score, then more from SpaceXAI (xAI).
| # | model | score | where | ||
|---|---|---|---|---|---|
| 216 |  Gemini 2.5 Flash | 62±1 | api $0.85/M | ||
| 217 |  MiniMax-M1 (40k) | 61±1 | comparison only | ||
| 218 | Grok 4 Fast | 61±2 | comparison only | ||
| 219 |  GLM-4.5V | 61±2 | api $0.90/M +local | ||
| 220 |  Llama 3.1 Nemotron Ultra 253B v1 | 61±2 | local | ||
| more from SpaceXAI (xAI) | |||||
| 30 | Grok 4.6 | 77±1 | api $3.00/M | ||
| 42 | Grok 4.5 | 75±1 | api $3.00/M | ||
| 47 | Grok 4.7 | 75±1 | api $3.00/M | ||
$ what it's good at
agent tasks▁▁▂▆█▆▅▄▄▃▂▁ #128 of 340 * · 7.0% of its score · 1 test
Beats o3-pro, o3, GPT-5 mini here, though they rank higher overall. Trails SmolLM3 3B, Jamba 1.5 Large here, though they rank lower overall.
| test | result | vs the others in its run | counts |
|---|---|---|---|
| τ³-bench banking_knowledge (v1.0.1 grading) | 15.7% ✎ | ▃▃▃▃·▆▃▂█▅▅▂ beats 4/25 | 7.0% |
coding▁▁▁▃▅▇█▇▅▄▃▁ #240 of 524 * · 35% of its score · 1 test
Trails QwQ-32B-Preview, Ministral 3 3B Reasoning 2512, DeepSeek-Coder-V2 Instruct (0724) here, though they rank lower overall.
| test | result | vs the others in its run | counts |
|---|---|---|---|
| Arena WebDev (Code Arena) | 1158.9 ✎ | ▁▃▄▃▅█▄▇▆▄▁▁ beats 1/119 | 35% |
chat & writing▁▁▁▃▅▄▆█▇▇▇▃ #120 of 463 * · 20% of its score · 1 test
Beats LongCat 2.0 (1.6T-A48B), GPT-5.4 nano, Step 3.5 Flash (196B-A11B) here, though they rank higher overall.
| test | result | vs the others in its run | counts |
|---|---|---|---|
| LMArena (text) | 1404.9 ✎ | ▁▁▁▂▃▃▃▆▆▇█▃ beats 250/379 | 20% |
vision & documents▁▁▃▃▂▃▅▄▇▇█▃ #154 of 182 * · 11% of its score · 2 tests
Trails Qwen3.5 2B, Qwen3-VL 2B Thinking, Molmo 72B (0924) here, though they rank lower overall.
| test | result | vs the others in its run | counts |
|---|---|---|---|
| CharXiv (reasoning questions) | 33.5% ✎ | ▄·····▄▄·▄·█ beats 0/5 | 7.0% |
| CharXiv descriptive questions | 68.2% ✎ | ▄····▄···█▄▄ beats 0/5 | 4.3% |
web search▂·▂▇▅█▃▄▁▂▁▁ #59 of 92 * · 8.1% of its score · 1 test
Beats Kimi K2.6, GPT-5 Pro, DeepSeek V4 Pro (Preview, Apr 2026) here, though they rank higher overall. Trails GPT-4o mini (2024-07-18), Mistral Small 3.1 24B (2503), Llama 4 Scout (109B-A17B) here, though they rank lower overall.
| test | result | vs the others in its run | counts |
|---|---|---|---|
| Arena search systems (style controlled) | 1160.3 ✎ | ▁·▁▁▂▃▂█▅▂▁▁ beats 6/30 | 8.1% |
other tests3 not in a task · 17% of its score
| test | result | vs the others in its run | counts |
|---|---|---|---|
| ARC-AGI-1 public evaluation set (ARC Prize model_baseline, pass@2) | 58.1% ✎ | ▁·▇▃▁·█▇▁·▇▃ beats 10/27 | 14% |
| ARC-AGI-2 public evaluation set (ARC Prize model_baseline, pass@2) | 4.6% ✎ | █··▂▅▁▃▁▃·▁▃ beats 9/24 | 3.5% |
| Vectara hallucination leaderboard (HHEM-2.3, factual consistency rate) | 79.8% ✎ | ▁▃▁▂▅█▂▃▄▃▁▁ beats 4/98 | 0.0% |
$ where to run it
No verified API or download recorded yet · add one
$ how the score was worked out · LMArena leaderboard dataset 63%, ARC Prize Foundation model_ba… 17%, ServiceNow-AI/Apriel-1.6-15b-… 11% and 2 more
Each result implies a score on its own, after adjusting for how hard its test is and who ran it. The score is their weighted average, with a small pull toward the middle for thinly tested models. Each test's weight is its counts column above; here it's added up by who ran the tests. How it works
| source | results | share of the score |
|---|---|---|
 LMArena leaderboard dataset | 3 | 63% |
 ARC Prize Foundation model_baseline outputs on ARC-AGI-1/2 public evaluation sets | 2 | 17% |
| ServiceNow-AI/Apriel-1.6-15b-Thinker model card | 2 | 11% |
 τ²/τ³-bench leaderboard: Sierra's own evaluation runs | 1 | 7% |
| Vectara hallucination leaderboard (vectara/hallucination-leaderboard repository) | 1 | 0% |
| prior · pull toward the middle | 1% | |
| weighted average | = 61.3 |
$ improve this page
Use a short GitHub form or edit the data through a pull request. Every number needs a source; here's which ones we can use.
data updated · sources and licences · json