Inference
EN

Gemini 4 Carbon vs GPT-6 Sol

Change models
Gemini 4 CarbonGoogle
GPT-6 SolOpenAI
IndexAll categories, out of 100.
82.7 ±5.1 best
72.8 ±2.2
GapIndex points below the top.
Top
−9.9
CoverageIndex weight that has results.
80%
100% best
SpeedOutput tokens per second.
31/s
61/s best
EloLMArena rating and rank.
1538 best
1397
ReleasedMonth of release.
2026-10
2026-09
Input / 1MUS dollars per 1M input tokens.
$2.00
$2.00
Output / 1MUS dollars per 1M output tokens.
$10.00
$10.00
Batch output / 1MBatch price per 1M output tokens.
$5.00
$5.00
1K chats2K tokens in, 500 out, each.
$9.00
$9.00
1K documents20K tokens in, 1K out, each.
$50.00
$50.00
1K agent tasks200K tokens in, 20K out, each.
$600.00
$600.00
ContextMaximum tokens in one request.
1.05M
1.05M
AgenticMulti-step tasks with tools.
83.2 best
68.6
CodingCode writing and repair.
80.5 best
73.9
ReasoningLogic problems and puzzles.
83.9 best
76.1
MultimodalTasks with images and text.
—
70.1
KnowledgeFacts and expert knowledge.
79.5 best
71.9
MultilingualTasks in many languages.
—
65.5
InstructionTasks with strict prompt rules.
—
65.3
MathMath problems.
77.2 best
75.8

Benchmarks

27 shared
Only Gemini 4 Carbon: 12Only GPT-6 Sol: 67
Gemini 4 Carbon
GPT-6 Sol
Artificial Analysis Long Context ReasoningReasoning
79.3%52 / 217
83.7% best10 / 217
GPQA diamond · Epoch AIKnowledge
96.5% best1 / 217
94.3%10 / 217
OTIS Mock AIME 2024-2025Math
100.0%1 / 200
100.0%1 / 200
Artificial Analysis Omniscience AccuracyKnowledge
58.3% best9 / 164
54.5%13 / 164
ARC-AGI-2 (semi-private)Reasoning
92.1% best5 / 131
89.6%10 / 131
ARC-AGI-1 (semi-private)Reasoning
97.5% best5 / 126
95.5%17 / 126
Vibe Code Bench v1.1Coding
89.1% best9 / 123
87.8%12 / 123
ARC-AGI-2 (public eval)Reasoning
97.9% best3 / 122
90.0%17 / 122
ARC-AGI-1 (public eval)Reasoning
98.5% best6 / 116
97.8%14 / 116
SimpleQA VerifiedKnowledge
77.7% best2 / 91
60.7%17 / 91
Chess PuzzlesReasoning
78.0% best2 / 90
56.0%15 / 90
FrontierMath-Tiers-1-3-v2-PrivateMath
90.5% best6 / 90
89.8%8 / 90
Legal Research BenchLegal
55.8% best1 / 90
28.8%57 / 90
Tax Agent BenchFinance
51.8% best1 / 82
15.0%62 / 82
Mystery Game PuzzlesReasoning
73.0% best5 / 80
56.0%12 / 80
FrontierMath-Tier-4-v2-PrivateMath
95.1% best4 / 72
90.0%8 / 72
Terminal-Bench 4.0Agentic
63.1% best4 / 61
44.4%13 / 61
Agent Arena steerabilityAgentic
19.1 best2 / 54
10.111 / 54
Agent Arena task outcomeAgentic
18.4 best4 / 54
4.517 / 54
Furniture AssemblyReasoning
77.5% best6 / 37
58.3%11 / 37