GPT-5.2 vs Claude Fable 5.1 vs Qwen3.8 Max (0902)

• GPT-5.2
• Claude Fable 5.1
• Qwen3.8 Max (0902)

Add a model

Add:Gemini 4 ArgonMuse Spark 1.1?Inkling Small 2DeepSeek V4P1 Flash

Head to head on 14 boards: Claude Fable 5.1 places best on 9, Qwen3.8 Max (0902) places best on 5 and GPT-5.2 places best on 0. All 3 ranked on 6 of 18 boards.

Two at a time on this screen. Tap one to swap it in.

GPT-5.2Claude Fable 5.1Qwen3.8 Max (0902)

Board	Read	GPT-5.2	Claude Fable 5.1	Qwen3.8 Max (0902)
Chat1
Text ArenaArena ratingtodaytoday	Nov 9todaytoday	#291,476	#61,504best of the picked	#201,483
Reasoning7
LiveBenchGlobal averagetodaytoday	Nov 9todaytoday	#3874.6%	#183.4%best of the picked	#1678.5%
Epoch ECIECItodaytoday	Nov 9todaytoday	#45154.4	#5166.9best of the picked	#27157.4
HLEAccuracytodaytoday	Nov 9todaytoday	#1427.8%	#246.5%best of the picked
ARC-AGI-2Scoretodaytoday	Nov 9todaytoday	#4152.9%	#790.0%best of the picked
AA IntelligenceIndextodaytoday	Nov 9todaytoday		#950.0best of the picked	#1533.7
GPQA DiamondAccuracytodaytoday	Nov 9todaytoday	#3191.4%		#2692.7%best of the picked
MMLU-ProAccuracy · staletodaytoday	Nov 9todaytoday	#1087.4%
Coding3
WebDev ArenaArena ratingtodaytoday	Nov 9todaytoday	#781,417	#51,746best of the picked	#121,677
SWE-Bench ProResolvedtodaytoday	Nov 9todaytoday		#292.2%
SWE-bench VerifiedResolved · staletodaytoday	Nov 9todaytoday	#672.8%
Agents and tools3
Search ArenaArena ratingtodaytoday	Nov 9todaytoday	#331,161	#71,227	#61,228best of the picked
MCP AtlasPass ratetodaytoday	Nov 9todaytoday	#2767.6%	#387.2%best of the picked
BFCLOverall accuracy · staletodaytoday	Nov 9todaytoday	#1455.9%
Vision2
Vision ArenaArena ratingtodaytoday	Nov 9todaytoday	#251,278	#131,288	#21,300best of the picked
Document ArenaArena ratingtodaytoday	Nov 9todaytoday	#371,425	#21,499best of the picked
Speed1
AA SpeedOutput speedtodaytoday	Nov 9todaytoday		#1422 tok/s	#260 tok/sbest of the picked
Price1
AA PriceBlended pricetodaytoday	Nov 9todaytoday		#17$20.0	#8$3.00best of the picked

Green marks the best placing among the picked where two or more are ranked; ties are marked for each. Ranks are each board's own order (lower is better), scores as the board published them. Dash: not on that board.