Comparing 2 models

Gemini 4 Carbon vs GPT-6.1 Sol

Benchmarks through

Verdict

Gemini 4 Carbon places higher on 18 of the 19 tasks where both are ranked. GPT-6.1 Sol accepts 1.0x more context.

  • Choose Gemini 4 Carbon for Document Summarization (#1), Logic & Problem Solving (#1), a General Assistant (#1).
  • Choose GPT-6.1 Sol for Tool & Function Calling (#5).

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 1Gemini 4 Carbon
Model 2GPT-6.1 Sol
Best for
a General Assistant (#1)
Image Understanding (#2)
Provider
google
openai
Context
1.0M
1.1MBest
Input / 1M
$2Best
$2Best
Output / 1M
$10Best
$10Best
Tool use
yesBest
yesBest
Vision
yesBest
yesBest
Reasoning
yesBest
yesBest
Open-weight
—
—
Released
Nov 10, 2026
Sep 29, 2026

Measured head to head

Benchmarks both models were run on

EvaluationGemini 4 CarbonGPT-6.1 SolSource
Design Arena Coding1348 Elo1362 Elosource · 2026-11-15
Design Arena UI1347 Elo1370 Elosource · 2026-11-15
Design Arena Website1319 Elo1315 Elosource · 2026-11-15
FACTS Grounding77.2%78%source · 2026-10-29
FrontierMath Tiers 1-3 (Epoch AI)90.5% ± 1.793.7% ± 1.4source · 2026-11-02
GPQA Diamond (Epoch AI)96.5% ± 1.295.4% ± 1.4source · 2026-11-02
LMArena Agent1712.5source · 2026-11-11
LMArena Agent Steerability19.113.4source · 2026-11-11
LMArena Agent Task Outcome18.411.7source · 2026-11-11
LMArena Agent Tool Hallucination0.51source · 2026-11-11
LMArena Creative Writing1533 Elo1420 Elosource · 2026-11-11
LMArena Instruction Following1539 Elo1463 Elosource · 2026-11-11
LMArena Long Query1549 Elo1459 Elosource · 2026-11-11
LMArena Text1526 Elo1481 Elosource · 2026-11-11
LMArena Text Factuality1521 Elo1468 Elosource · 2026-11-11
LMArena WebDev1768 Elo1743 Elosource · 2026-11-11
OTIS Mock AIME (Epoch AI)100% ± 0100% ± 0source · 2026-11-02
SimpleQA Verified (Epoch AI)77.7% ± 1.373.9% ± 1.4source · 2026-11-02
Terminal-Bench (4.1)64.5%60.9%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.