DeepSeek-V3.2 vs GPT-6 Astra vs Claude Fable 5.1

• DeepSeek-V3.2
• GPT-6 Astra
• Claude Fable 5.1

Add a model

Add:Gemini 4 ArgonMuse Spark 1.1Qwen3.8 Max (0902)?Inkling Small 2

Head to head on 13 boards: GPT-6 Astra places best on 7, Claude Fable 5.1 places best on 7 and DeepSeek-V3.2 places best on 0 (a tie counts for each). All 3 ranked on 2 of 19 boards.

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

DeepSeek-V3.2GPT-6 AstraClaude Fable 5.1

Board	Read	DeepSeek-V3.2	GPT-6 Astra	Claude Fable 5.1
Chat1
Text ArenaArena ratingtodaytoday	Nov 9todaytoday		#311,475	#61,504best of the picked
Reasoning7
Epoch ECIECItodaytoday	Nov 9todaytoday	#82146.2	#1169.1best of the picked	#5166.9
ARC-AGI-2Scoretodaytoday	Nov 9todaytoday	#684.0%	#195.0%best of the picked	#790.0%
HLEAccuracytodaytoday	Nov 9todaytoday		#154.8%best of the picked	#246.5%
LiveBenchGlobal averagetodaytoday	Nov 9todaytoday		#482.2%	#183.4%best of the picked
AA IntelligenceIndextodaytoday	Nov 9todaytoday		#356.9best of the picked	#950.0
GPQA DiamondAccuracytodaytoday	Nov 9todaytoday		#295.8%
MMLU-ProAccuracy · staletodaytoday	Nov 9todaytoday	#3385.0%
Coding4
SWE-Bench ProResolvedtodaytoday	Nov 9todaytoday		#490.2%	#292.2%best of the picked
WebDev ArenaArena ratingtodaytoday	Nov 9todaytoday		#31,770best of the picked	#51,746
Aider PolyglotPass rate · staletodaytoday	Nov 9todaytoday	#874.2%
SWE-bench VerifiedResolved · staletodaytoday	Nov 9todaytoday	#1170.0%
Agents and tools3
Search ArenaArena ratingtodaytoday	Nov 9todaytoday		#31,245best of the picked	#71,227
MCP AtlasPass ratetodaytoday	Nov 9todaytoday			#387.2%
BFCLOverall accuracy · staletodaytoday	Nov 9todaytoday	#1256.7%
Vision2
Document ArenaArena ratingtodaytoday	Nov 9todaytoday		#111,477	#21,499best of the picked
Vision ArenaArena ratingtodaytoday	Nov 9todaytoday		#171,285	#131,288best of the picked
Speed1
AA SpeedOutput speedtodaytoday	Nov 9todaytoday		#1621 tok/s	#1422 tok/sbest of the picked
Price1
AA PriceBlended pricetodaytoday	Nov 9todaytoday		#17$20.0best of the picked	#17$20.0best 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.