Inference
EN

Grok 4.7 vs Step 5 Preview

Change models
Grok 4.7xAI
Step 5 PreviewStepFun
IndexAll categories, out of 100.
69.1 ±3.7 best
68.4 ±3.7
GapIndex points below the top.
Top
−0.7 · no clear gap
CoverageIndex weight that has results.
85%
90% best
SpeedOutput tokens per second.
51/s
105/s best
EloLMArena rating and rank.
1398
1443 best
ReleasedMonth of release.
2026-09
2026-09
Input / 1MUS dollars per 1M input tokens.
$2.00
$1.00 best
Output / 1MUS dollars per 1M output tokens.
$6.00
$2.70 best
Batch output / 1MBatch price per 1M output tokens.
$4.80
—
1K chats2K tokens in, 500 out, each.
$7.00
$3.35 best
1K documents20K tokens in, 1K out, each.
$46.00
$22.70 best
1K agent tasks200K tokens in, 20K out, each.
$520.00
$254.00 best
ContextMaximum tokens in one request.
500K
1M best
AgenticMulti-step tasks with tools.
71.9 best
68.8
CodingCode writing and repair.
69.9
71.4 best
ReasoningLogic problems and puzzles.
64.5
67.4 best
MultimodalTasks with images and text.
—
62.2
KnowledgeFacts and expert knowledge.
66.5 best
63.5
MultilingualTasks in many languages.
—
—
InstructionTasks with strict prompt rules.
78.9
—
MathMath problems.
61.8 best
61.7

Benchmarks

41 shared
Only Grok 4.7: 43Only Step 5 Preview: 21
Grok 4.7
Step 5 Preview
Artificial Analysis Humanity's Last ExamKnowledge
43.1%32 / 222
46.5% best23 / 222
Artificial Analysis Long Context ReasoningReasoning
76.7%84 / 217
88.3% best2 / 217
Critical Physics TasksReasoning
17.7%37 / 214
20.9% best26 / 214
Artificial Analysis Omniscience AccuracyKnowledge
42.7% best42 / 164
36.4%64 / 164
GDPval-AA normalizedAgentic
60.7% best8 / 133
54.3%24 / 133
Vibe Code Bench v1.1Coding
86.2% best13 / 123
70.4%48 / 123
Artificial Analysis SciCodeCoding
57.4%19 / 114
58.9% best13 / 114
Legal Research BenchLegal
47.1% best16 / 90
40.4%35 / 90
Tax Agent BenchFinance
35.8% best16 / 82
27.3%39 / 82
LiveBench Data AnalysisReasoning
76.9% best30 / 80
63.5%69 / 80
LiveBench ReasoningReasoning
82.7%51 / 80
83.8% best44 / 80
Finance Agent (v2)Finance
52.3% best36 / 79
50.7%42 / 79
Code MigrationCoding
44.8% best19 / 77
39.0%31 / 77
EMBFinance
67.0% best20 / 75
57.1%41 / 75
Terminal-Bench 4.0Agentic
28.8%25 / 61
31.8% best23 / 61
DeepSWEAgentic
71.0% best21 / 59
67.7%27 / 59
Agent Arena steerabilityAgentic
−2.335 / 54
−0.9 best30 / 54
Agent Arena task outcomeAgentic
5.1 best14 / 54
−1.535 / 54
ProofBench v1.1Math
26.0%40 / 54
42.0% best37 / 54
LiveBench Coding 2.0Coding
84.5%24 / 52
85.6% best22 / 52