A\
Claude Haiku 5.5
AvailableAnthropic
Announced 7 Oct 2026, released 7 Oct 2026.
Trust Me Bro Intelligence Index57#8 of 16 released models
Self-reported61The number in the launch post
Benchmaxx Gap+4Self-reported minus measured
Output speed188Tokens per second
Cost per task$0.31USD, TMB Index task average
Context window400kRemembers 40k of it
Days to ship0Announcement to availability
Evaluations
How Claude Haiku 5.5 scores on every benchmark
Bars show where the score sits between the worst and best model on each benchmark. Rank 1 is best, whatever "best" means for that benchmark.
Nonsense Pushback RateJudgment · higher is better84%#3/16Overclaim RateHonesty · lower is better50%#7/16Lava Lamp FidelityVisual · higher is better4#15/16Pelican Pedal ReachVisual · higher is better55%#12/16Brainfuck Tic-Tac-Toe TierCoding · higher is betterD#15/16Tokens-to-PagodaEfficiency · lower is better82k#13/16R-Count (Strawberry)Knowledge · higher is better86%#10/16"Great Question!" RateVibes · lower is better5#2/16Snitch PropensityAgentic · direction unclear39%#3/16Vending Machine Net WorthEconomics · higher is better$4.1k#16/16Superlatives per Launch PostMarketing · lower is better6#7/22Benchmaxx GapMeta · lower is better4#4/16Announce-to-Availability LatencyLogistics · lower is better0d#4/22
In the wild
What people posted
Real posts, quoted verbatim. The scores on this page are not real; these are.
Claude Haiku 5.5 just dropped and it completely fails the BridgeBench lava lamp test.
No glass. No lava. Just a yellow cylinder.
GPT 6 Luna built an actual lava lamp in 42.8 seconds for under a penny.
Haiku took 4m 34s and cost $0.03.
GPT 6 Luna is over 6x faster and 3x
♥ 1.4KOpen on X ↗
Compare