Comparing 2 models

Claude Sonnet 5.5 vs Gemini 4 Carbon

Benchmarks through

Verdict

Gemini 4 Carbon places higher on 11 of the 14 tasks where both are ranked. Gemini 4 Carbon accepts 1.0x more context.

  • Choose Claude Sonnet 5.5 for Coding (#3), Frontend & UI (#3), Tool & Function Calling (#6).
  • Choose Gemini 4 Carbon for RAG & Long Context (#1), Math & Reasoning (#2), Competition Math (#2).

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 4 Carbon
Best for
Code Completion (#1)
a General Assistant (#1)
Provider
anthropic
google
Context
1M
1.0MBest
Input / 1M
$2Best
$2Best
Output / 1M
$10Best
$10Best
Tool use
yesBest
yesBest
Vision
yesBest
yesBest
Reasoning
yesBest
yesBest
Open-weight
—
—
Released
Sep 28, 2026
Nov 10, 2026

Measured head to head

Benchmarks both models were run on

EvaluationClaude Sonnet 5.5Gemini 4 CarbonSource
Design Arena Coding1345 Elo1348 Elosource · 2026-11-15
Design Arena UI1389 Elo1347 Elosource · 2026-11-15
Design Arena Website1309 Elo1319 Elosource · 2026-11-15
FACTS Grounding62.5%77.2%source · 2026-11-12
FrontierMath Tiers 1-3 (Epoch AI)88.8% ± 1.990.5% ± 1.7source · 2026-09-29
GPQA Diamond (Epoch AI)95.6% ± 1.496.5% ± 1.2source · 2026-09-29
LMArena Agent1617source · 2026-11-11
LMArena Agent Steerability14.519.1source · 2026-11-11
LMArena Agent Task Outcome19.518.4source · 2026-11-11
LMArena Agent Tool Hallucination0.90.5source · 2026-11-11
LMArena Creative Writing1466 Elo1533 Elosource · 2026-11-11
LMArena Instruction Following1487 Elo1539 Elosource · 2026-11-11
LMArena Long Query1487 Elo1549 Elosource · 2026-11-11
LMArena Text1473 Elo1526 Elosource · 2026-11-11
LMArena Text Factuality1472 Elo1521 Elosource · 2026-11-11
LMArena WebDev1772 Elo1768 Elosource · 2026-11-11
OTIS Mock AIME (Epoch AI)100% ± 0100% ± 0source · 2026-09-29
SimpleQA Verified (Epoch AI)46.5% ± 1.677.7% ± 1.3source · 2026-09-29
Terminal-Bench (4.1)63.9%64.5%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.