Comparing 2 models

Claude Sonnet 5.5 vs Gemini 3.8 Flash

Benchmarks through

Verdict

Gemini 3.8 Flash places higher on 1 of the 1 tasks where both are ranked. Gemini 3.8 Flash costs 2.4x less per input token. Gemini 3.8 Flash accepts 1.0x more context.

  • Choose Gemini 3.8 Flash for RAG & Long Context (#6).

Built from the published task rankings below; ranks come from each task's own weighted benchmarks, so a win on one task says nothing about another.

metric
Model 1Claude Sonnet 5.5
Model 2Gemini 3.8 Flash
Best for
Code Completion (#1)
Business Writing (#1)
Provider
anthropic
google
Context
1M
1.0MBest
Input / 1M
$2
$0.83Best
Output / 1M
$10
$4.13Best
Tool use
yesBest
yesBest
Vision
yesBest
yesBest
Reasoning
yesBest
yesBest
Open-weight
—
—
Released
Sep 28, 2026
Sep 2, 2026

Measured head to head

Benchmarks both models were run on

EvaluationClaude Sonnet 5.5Gemini 3.8 FlashSource
Design Arena Coding1345 Elo1292 Elosource · 2026-11-15
Design Arena UI1389 Elo1303 Elosource · 2026-11-15
Design Arena Web-app Agent1295 Elo1213 Elosource · 2026-11-15
Design Arena Website1309 Elo1291 Elosource · 2026-11-15
FACTS Grounding62.5%68.2%source · 2026-11-12
FrontierMath Tiers 1-3 (Epoch AI)88.8% ± 1.968.4% ± 2.8source · 2026-09-29
GPQA Diamond (Epoch AI)95.6% ± 1.495.4% ± 1.4source · 2026-09-29
LiveBench Agentic Coding56.3%54.2%source · 2026-10-18
LiveBench Coding91.4%72.5%source · 2026-10-18
LiveBench Data Analysis59.5%54.0%source · 2026-10-18
LiveBench Instruction Following56.8%81.4%source · 2026-10-18
LiveBench Language78.0%87.8%source · 2026-10-18
LiveBench Math96.1%91.6%source · 2026-10-18
LiveBench Reasoning91.6%89.3%source · 2026-10-18
LMArena Agent162.3source · 2026-11-11
LMArena Agent Bash Recovery15-0.7source · 2026-10-08
LMArena Agent Steerability14.5-1.3source · 2026-11-11
LMArena Agent Task Outcome19.51.3source · 2026-11-11
LMArena Agent Tool Hallucination0.90.4source · 2026-11-11
LMArena Creative Writing1466 Elo1493 Elosource · 2026-11-11
LMArena Instruction Following1487 Elo1493 Elosource · 2026-11-11
LMArena Long Query1487 Elo1508 Elosource · 2026-11-11
LMArena Math1483 Elo1525 Elosource · 2026-11-11
LMArena Text1473 Elo1497 Elosource · 2026-11-11
LMArena Text Factuality1472 Elo1470 Elosource · 2026-11-11
LMArena Vision1263 Elo1293 Elosource · 2026-11-09
LMArena Vision Diagrams1308 Elo1342 Elosource · 2026-11-09
LMArena Vision OCR1281 Elo1315 Elosource · 2026-11-09
LMArena WebDev1772 Elo1583 Elosource · 2026-11-11
LMArena WebDev Image-to-Web1744 Elo1594 Elosource · 2026-11-10
OTIS Mock AIME (Epoch AI)100% ± 098.9% ± 0.9source · 2026-09-29
SimpleQA Verified (Epoch AI)46.5% ± 1.669.7% ± 1.5source · 2026-09-29
Terminal-Bench (4.1)63.9%18.2%source · 2026-11-15

Values are as published by each evaluator, in its own units; rows from different evaluators are not on a common scale.

task evidence

How they place when the question changes

Lower ranks are better. “Best” marks the strongest option in this shortlist for that task.