Inference
EN

Claude Sonnet 5.5 vs Gemini 3.8 Flash

Change models
Claude Sonnet 5.5Anthropic
Gemini 3.8 FlashGoogle
IndexAll categories, out of 100.
77.2 ±2.6 best
67.8 ±2.5
GapIndex points below the top.
Top
−9.4
CoverageIndex weight that has results.
95%
95%
SpeedOutput tokens per second.
107/s best
68/s
EloLMArena rating and rank.
1469
1500 best
ReleasedMonth of release.
2026-09
2026-09
Input / 1MUS dollars per 1M input tokens.
$2.00
$0.825 best
Output / 1MUS dollars per 1M output tokens.
$10.00
$4.13 best
Batch output / 1MBatch price per 1M output tokens.
$5.00
$1.88 best
1K chats2K tokens in, 500 out, each.
$9.00
$3.71 best
1K documents20K tokens in, 1K out, each.
$50.00
$20.63 best
1K agent tasks200K tokens in, 20K out, each.
$600.00
$247.50 best
ContextMaximum tokens in one request.
1M
1.05M best
AgenticMulti-step tasks with tools.
77.5 best
65.3
CodingCode writing and repair.
80.8 best
66.2
ReasoningLogic problems and puzzles.
77.8 best
68.3
MultimodalTasks with images and text.
74.4 best
67.1
KnowledgeFacts and expert knowledge.
71.3 best
70.9
MultilingualTasks in many languages.
—
—
InstructionTasks with strict prompt rules.
41.5
91.3 best
MathMath problems.
76.1 best
64.3

Benchmarks

76 shared
Only Claude Sonnet 5.5: 52Only Gemini 3.8 Flash: 30
Claude Sonnet 5.5
Gemini 3.8 Flash
Artificial Analysis Humanity's Last ExamKnowledge
55.0% best6 / 222
47.8%20 / 222
Artificial Analysis Long Context ReasoningReasoning
82.7% best23 / 217
81.3%34 / 217
GPQA diamond · Epoch AIKnowledge
95.6% best4 / 217
95.4%5 / 217
Critical Physics TasksReasoning
31.4% best7 / 214
18.3%34 / 214
OTIS Mock AIME 2024-2025Math
100.0% best1 / 200
98.9%22 / 200
Artificial Analysis Omniscience AccuracyKnowledge
51.5% best17 / 164
50.5%19 / 164
GDPval-AA normalizedAgentic
66.9% best4 / 133
46.8%34 / 133
ARC-AGI-2 (semi-private)Reasoning
83.9%17 / 131
89.2% best11 / 131
ARC-AGI-1 (semi-private)Reasoning
96.5%11 / 126
98.5% best1 / 126
Vibe Code Bench v1.1Coding
92.4% best2 / 123
78.7%34 / 123
ARC-AGI-2 (public eval)Reasoning
90.1%16 / 122
92.5% best11 / 122
ARC-AGI-1 (public eval)Reasoning
97.8%14 / 116
98.0% best11 / 116
Artificial Analysis SciCodeCoding
61.0% best6 / 114
56.6%23 / 114
MedScribeHealthcare
91.1% best3 / 109
84.5%34 / 109
MedCodeHealthcare
52.9% best12 / 106
48.1%31 / 106
SimpleQA VerifiedKnowledge
46.5%39 / 91
69.7% best10 / 91
Chess PuzzlesReasoning
59.0% best10 / 90
58.0%14 / 90
FrontierMath-Tiers-1-3-v2-PrivateMath
88.8% best10 / 90
68.4%32 / 90
Legal Research BenchLegal
48.1% best12 / 90
38.9%38 / 90
Tax Agent BenchFinance
42.0% best10 / 82
32.1%23 / 82