
None
None
Long Story Video Skill
Turn your idea into 10s–10min video—AI writes scripts, prompts, and generates footage automatically.

Ads Video Skill
Generate professional ads and sales videos—AI auto-generates scripts, prompts, and footage.
3D Science Explainer Video Skill
Convert scientific concepts into stunning 3D explain animations
Feedback
freeTrialImage.bannerPity
freeTrialImage.upgradeUnlock
- ✓freeTrialImage.benefitHd
- ✓freeTrialImage.benefitWatermark
- ✓freeTrialImage.benefitUnlimited
Haiku 5.5 Benchmark
Haiku 5.5 Benchmark explained: launch scores, price tiers and published limits, each tied to its source and date — no hands-on lab test claimed.
All Tools
Discover our comprehensive AI-powered animation toolkit
MiniMax H3
MiniMax H3 AI Video Generator
Seedance 2.5
The Future of AI Video Is Here.

Seedance 2.0
The Future of AI Video Is Here.

Seedance 2.0
The Future of AI Video Is Here.

AI Multi-Scene Shorts Generator
Create viral AI Shorts instantly

3D Science Video
Create 3D science videos easily

Suno AI Music Generator
Create Professional Music with AI

Free GPT Image 2.5
Next-generation image creator
What Makes the Haiku 5.5 Benchmark Worth a Close Look
Launch-day scores, price bands and per-effort costs, each kept alongside the configuration it came from and the date it was retrieved.
- What Launch Scores Can and Cannot Tell YouElo, partial-credit, coding and accuracy numbers sit on incompatible scales, so averaging them together would be misleading.
- Small-Model Economics for High-Volume WorkDesigned for extraction, classification and summarisation jobs that repeat thousands of times a day.
- Two Caveats That Reshape the Score TablesThe low rate stops at 100K tokens, and the 72.4% computer-use figure is partial credit rather than a pass rate.
Reading the Haiku 5.5 Benchmark in Three Steps
A short workflow for turning published Haiku 5.5 numbers into a real workload decision.
Haiku 5.5 Benchmark: The Findings That Matter
Published specs, price bands, coding outcomes, computer-use scores and third-party index data for Claude Haiku 5.5, each with its caveat attached.
Context Window and Output Ceiling
Text-plus-image input, text output, adaptive thinking on. A one-million-token context and a 128,000-token output cap overshadow Haiku 4.5's 200,000 and 64,000.
Two Price Bands Set by Prompt Length
Past 100,000 prompt tokens, input, output and caching each move to a band five times the cheap one. OpenAI's base Standard band reaches 272,000 input tokens.
Terminal-Bench 4.0: Finished Task Runs
Haiku's 39.2% launch result clearly beats Luna but stays below the published Sol rows. Runs get an eight-hour agent timeout, and the 2.1 or 3.0 figures come from an older experiment.
FrontierCode: Mergeability Versus Effort Cost
Grading weighs correctness, tests, scope, style and repo conventions. Haiku reaches 46.4% at max effort and 41.6% at medium for $0.13 a rollout, while moving from xhigh to max doubles spend for half a point.
OSWorld 2.1: Partial Credit or Strict Pass
Eighty-two offline tasks, five attempts each, 1080p screenshots and a 500-step ceiling shape the run. For workflows that must finish, the 37.1% strict rate is the number to trust.
Third-Party Index Scores and Tool Gains
Artificial Analysis lists Haiku across all five efforts on Intelligence Index v4.3.2 with rounded task costs. HLE reads 45.9% without tools and 57.4% with them — a lift from the surrounding system.
Haiku 5.5 Benchmark: Frequently Asked Questions
Straight answers on Claude Haiku 5.5 results, pricing thresholds and published limits.
Where does Haiku 5.5 stand next to GPT-6 Luna and GPT-6 Sol?
On knowledge work, screenshot tasks and narrow coding, Haiku and Luna trade blows. Sol and Sol 6.1 charge more and justify it on coding, while Sonnet 5.5 still leads most rows.
Why doesn't the 72.4% computer-use figure count as a success rate?
The same run reports only 37.1% when every checkpoint has to be met. An agent can find and edit the right file, then save it in the wrong format or stop before the final action. Always add an end-state check.
At what point does Haiku 5.5 move into the higher price band?
Only prompts up to 100,000 tokens receive the cheapest rates; beyond that, every category costs five times as much. Anthropic's 90%, 50% and ~75% savings figures are measured against different baselines.
Could Haiku 5.5 run locally, and what GPU memory would that take?
The published specs list no parameter count, active-parameter figure, layer count or training compute. 'Small' describes a product tier, not an architecture, so local GPU estimates have nothing to stand on.
Does a higher reasoning effort always mean better results?
No. Sol 6.1 scores better at medium than at max, and Haiku's 0.2-point lead over Sonnet at max is too narrow to call a win. Extra effort adds tokens, latency and more chances to drift off task.
Why do identical models show different numbers across tables?
Sonnet reads 70.6% in Anthropic's launch table and 61.8% on the public leaderboard snapshot, where Haiku had no visible row. Artificial Analysis dropped its 'Default Fallback' label, yet Anthropic's migration guide still states that no server-side fallback exists.
Put the Haiku 5.5 Benchmark to Work on Your Own Tasks
Choose one workload, set the effort level, and measure Haiku 5.5 against the quality floor you actually need.
