Inference
EN

Claude Opus 5.5 vs Gemini 4 Carbon

Change models
Claude Opus 5.5Anthropic
Gemini 4 CarbonGoogle
IndexAll categories, out of 100.
78.6 ±2.4
82.7 ±5.1 best
GapIndex points below the top.
−4.1 · no clear gap
Top
CoverageIndex weight that has results.
95% best
80%
SpeedOutput tokens per second.
76/s best
31/s
EloLMArena rating and rank.
1513
1538 best
ReleasedMonth of release.
2026-09
2026-10
Input / 1MUS dollars per 1M input tokens.
$4.00
$2.00 best
Output / 1MUS dollars per 1M output tokens.
$20.00
$10.00 best
Batch output / 1MBatch price per 1M output tokens.
$10.00
$5.00 best
1K chats2K tokens in, 500 out, each.
$18.00
$9.00 best
1K documents20K tokens in, 1K out, each.
$100.00
$50.00 best
1K agent tasks200K tokens in, 20K out, each.
$1,200.00
$600.00 best
ContextMaximum tokens in one request.
1M
1.05M best
AgenticMulti-step tasks with tools.
74.1
83.2 best
CodingCode writing and repair.
81.8 best
80.5
ReasoningLogic problems and puzzles.
80.8
83.9 best
MultimodalTasks with images and text.
74.1
—
KnowledgeFacts and expert knowledge.
79.2
79.5 best
MultilingualTasks in many languages.
—
—
InstructionTasks with strict prompt rules.
59.6
—
MathMath problems.
77.7 best
77.2

Benchmarks

37 shared
Only Claude Opus 5.5: 120Only Gemini 4 Carbon: 2
Claude Opus 5.5
Gemini 4 Carbon
Artificial Analysis Long Context ReasoningReasoning
84.7% best6 / 217
79.3%52 / 217
GPQA diamond · Epoch AIKnowledge
90.6%38 / 217
96.5% best1 / 217
OTIS Mock AIME 2024-2025Math
100.0%1 / 200
100.0%1 / 200
Artificial Analysis Omniscience AccuracyKnowledge
65.5% best3 / 164
58.3%9 / 164
ARC-AGI-2 (semi-private)Reasoning
91.7%6 / 131
92.1% best5 / 131
ARC-AGI-1 (semi-private)Reasoning
97.5%5 / 126
97.5%5 / 126
Vibe Code Bench v1.1Coding
90.3% best6 / 123
89.1%9 / 123
ARC-AGI-2 (public eval)Reasoning
97.9%3 / 122
97.9%3 / 122
ARC-AGI-1 (public eval)Reasoning
98.3%8 / 116
98.5% best6 / 116
SimpleQA VerifiedKnowledge
72.2%7 / 91
77.7% best2 / 91
Chess PuzzlesReasoning
67.0%5 / 90
78.0% best2 / 90
FrontierMath-Tiers-1-3-v2-PrivateMath
91.2% best4 / 90
90.5%6 / 90
Legal Research BenchLegal
50.5%7 / 90
55.8% best1 / 90
Tax Agent BenchFinance
44.5%6 / 82
51.8% best1 / 82
Mystery Game PuzzlesReasoning
71.0%6 / 80
73.0% best5 / 80
FrontierMath-Tier-4-v2-PrivateMath
95.0%5 / 72
95.1% best4 / 72
Terminal-Bench 4.0Agentic
65.2% best1 / 61
63.1%4 / 61
Agent Arena steerabilityAgentic
18.63 / 54
19.1 best2 / 54
Agent Arena task outcomeAgentic
20.7 best2 / 54
18.44 / 54
Furniture AssemblyReasoning
83.3% best2 / 37
77.5%6 / 37