Inference
EN

GPT-5.6 Terra vs Step 5 Preview

Change models
GPT-5.6 TerraOpenAI
Step 5 PreviewStepFun
IndexAll categories, out of 100.
68.9 ±2.0 best
68.4 ±3.7
GapIndex points below the top.
Top
−0.5 · no clear gap
CoverageIndex weight that has results.
100% best
90%
SpeedOutput tokens per second.
66/s
105/s best
EloLMArena rating and rank.
1449 best
1443
ReleasedMonth of release.
2026-07
2026-09
Input / 1MUS dollars per 1M input tokens.
$2.00
$1.00 best
Output / 1MUS dollars per 1M output tokens.
$12.00
$2.70 best
Batch output / 1MBatch price per 1M output tokens.
$6.00
—
1K chats2K tokens in, 500 out, each.
$10.00
$3.35 best
1K documents20K tokens in, 1K out, each.
$52.00
$22.70 best
1K agent tasks200K tokens in, 20K out, each.
$640.00
$254.00 best
ContextMaximum tokens in one request.
1.05M best
1M
AgenticMulti-step tasks with tools.
68.2
68.8 best
CodingCode writing and repair.
68.4
71.4 best
ReasoningLogic problems and puzzles.
72.5 best
67.4
MultimodalTasks with images and text.
67.3 best
62.2
KnowledgeFacts and expert knowledge.
65.5 best
63.5
MultilingualTasks in many languages.
60.7
—
InstructionTasks with strict prompt rules.
56.4
—
MathMath problems.
72.8 best
61.7

Benchmarks

38 shared
Only GPT-5.6 Terra: 71Only Step 5 Preview: 24
GPT-5.6 Terra
Step 5 Preview
Artificial Analysis Humanity's Last ExamKnowledge
42.9%34 / 222
46.5% best23 / 222
Artificial Analysis Long Context ReasoningReasoning
83.0%18 / 217
88.3% best2 / 217
Critical Physics TasksReasoning
30.0% best10 / 214
20.9%26 / 214
Artificial Analysis Omniscience AccuracyKnowledge
46.8% best31 / 164
36.4%64 / 164
GDPval-AA normalizedAgentic
47.6%33 / 133
54.3% best24 / 133
Vibe Code Bench v1.1Coding
74.6% best44 / 123
70.4%48 / 123
Artificial Analysis SciCodeCoding
55.0%29 / 114
58.9% best13 / 114
Artificial Analysis MMMU-ProMultimodal
80.7% best17 / 113
76.4%44 / 113
Legal Research BenchLegal
41.3% best30 / 90
40.4%35 / 90
Terminal-Bench 2.1 · BenchLMAgentic
87.4% best14 / 86
85.0%20 / 86
GPQA Diamond · BenchLMKnowledge
92.9%13 / 84
93.5% best10 / 84
Tax Agent BenchFinance
30.0% best34 / 82
27.3%39 / 82
LiveBench Data AnalysisReasoning
79.3% best15 / 80
63.5%69 / 80
LiveBench ReasoningReasoning
90.6% best11 / 80
83.8%44 / 80
Finance Agent (v2)Finance
54.4% best27 / 79
50.7%42 / 79
Code MigrationCoding
47.8% best13 / 77
39.0%31 / 77
EMBFinance
66.2% best23 / 75
57.1%41 / 75
BrowseCompAgentic
87.5%14 / 63
88.7% best12 / 63
Terminal-Bench 4.0Agentic
22.7%32 / 61
31.8% best23 / 61
DeepSWEAgentic
69.6% best23 / 59
67.7%27 / 59