MMODELYST
Models/Grok 3 mini Reasoning
MDL

Grok 3 mini Reasoning

xAI

Verified against Artificial Analysis · Jul 13, 2026
Capability
78
percentile index
Price /M
$0.3 / $0.5
in / out
Context
Avg score
61.0
11 benchmarks
AA Intelligence
22.5
index
Speed
47 tok/s
TTFT 0.6s
Released
Feb 19, 2025
About

Grok 3 mini Reasoning is a proprietary model in the Grok family from xAI. Benchmarked on 11 evals, averaging 61.0.

Benchmark scores · 11
avg 61.0 — higher is better
MATH
99.2
τ²-bench
90.4
AIME
84.7
MMLU-Pro
82.8
GPQA Diamond
79.1
LiveCodeBench
69.6
AA Long-Context Reasoning
50.3
IFBench
45.9
SciCode
40.6
Terminal-Bench Hard
17.4
Humanity's Last Exam
11.1
Details
PublisherxAI
FamilyGrok
Parameters
License
WeightsClosed
ReleasedFeb 19, 2025
Performance & price
Output speed47 tok/s
Latency (TTFT)0.6s
First answer43s
Price in / out$0.3 / $0.5 per 1M
Blended price$0.35 / 1M
Math index84.7

Median across API providers, via Artificial Analysis. Blended = 3:1 input:output. For reasoning models, latency includes thinking time. Methodology →

Capability over time

Append-only ledger — every observed change to this model's numbers.

API providers
Artificial Analysis$0.5/M