Comparing 2 models

Claude Fable 5.2 vs Kimi K3

Benchmarks through

Verdict

Claude Fable 5.2 and Kimi K3 are not yet ranked on the same task, so compare the specifications below. Kimi K3 costs 41.7x less per input token. Kimi K3 accepts 1.0x more context. Only Kimi K3 has open weights for self-hosting.

  • Choose Kimi K3 for you need to self-host.

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 Fable 5.2
Model 2Kimi K3
Best for
Agentic Coding (#1)
JSON & Schema Output (#2)
Provider
anthropic
moonshotai
Context
1M
1.0MBest
Input / 1M
$8
$0.19Best
Output / 1M
$40
$16Best
Tool use
yesBest
yesBest
Vision
yesBest
yesBest
Reasoning
yesBest
yesBest
Open-weight
—
yesBest
Released
Nov 9, 2026
Jul 16, 2026

Measured head to head

Benchmarks both models were run on

EvaluationClaude Fable 5.2Kimi K3Source
Design Arena Coding1375 Elo1352 Elosource · 2026-11-15
Design Arena Full-stack Agent1400 Elo1289 Elosource · 2026-11-15
Design Arena UI1382 Elo1342 Elosource · 2026-11-15
Design Arena Website1352 Elo1336 Elosource · 2026-11-15
FrontierMath Tiers 1-3 (Epoch AI)95.1% ± 1.372.2% ± 2.7source · 2026-11-06
GPQA Diamond (Epoch AI)93.9% ± 1.693.1% ± 1.5source · 2026-11-06
LMArena Agent30.40.6source · 2026-11-11
LMArena Agent Steerability33-0.8source · 2026-11-11
LMArena Agent Task Outcome223.4source · 2026-11-11
LMArena Agent Tool Hallucination11source · 2026-11-11
LMArena WebDev1838 Elo1658 Elosource · 2026-11-11
OTIS Mock AIME (Epoch AI)99.7% ± 0.397.2% ± 1.1source · 2026-11-06
SimpleQA Verified (Epoch AI)94% ± 0.850.6% ± 1.6source · 2026-11-06
Terminal-Bench (4.1)65.8%23.0%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.