Models / claude-haiku-5-5

Claude Haiku 5.5

Anthropic's model, with closed weights. Forcebench has run it in 3 configurations on the API board, plus 1 on the lite subset. API is models reached through a hosted service; Local, models served on Forcebench’s own hardware.

On the API board

Models reached through a hosted service. Claude Haiku 5.5 is ranked by the highest-scoring of its 3 configurations: the 14th-highest score of 26 models, not clearly different from 5 other models. Scores are within 3 attempts (c@3) with 95% intervals; each configuration is compared with the highest-scoring one by whether their 95% intervals overlap. Every configuration on the board → Cost and speed →

76%
within 3 attempts (c@3), 95% CI 72–81 · medium · 14th-highest score of 26 models, not clearly different from 5 other models
2.8k
output tokens per task · median 1.7k, p90 7k
8.7 s
median per task, every attempt counted, including network and queueing · 90th percentile 44 s
0%
no answer within the 32,768-token budget
$0.0025
per task at the list price of 10 October 2026, every attempt counted
$0.0032
per solved task: the total cost over the tasks solved · cost and speed
Every configuration
Within 3 attempts (c@3) with 95% intervals, highest first. Bars whose intervals overlap are not clearly different.
  • 95% interval
medium Anthropichighest score
76%, 95% interval 72–81%
high Anthropic
76%, 95% interval 72–80%
low Anthropic
74%, 95% interval 69–78%
    Show the numbers
    Every configuration of Claude Haiku 5.5 on the API board
    ConfigurationRank all configsWithin 3 attempts c@3 · 95% CIAgainst the highest by 95% interval overlapTokens mean · median · p90No answerLatency median per task, every attempt countedCost per taskCost per solved task
    Anthropic · mediumhighest score22
    76%95% CI 72–81
    		—2.8k1.7k · 7k0%8.7 s$0.0025$0.0032
    Anthropic · high22
    76%95% CI 72–80
    		not clearly different4.5k2.5k · 11k<1%12 s$0.0039$0.0051
    Anthropic · low25
    74%95% CI 69–78
    		not clearly different2.2k1.4k · 4.8k0%7.5 s$0.0021$0.0028
    Suites, at its highest-scoring configuration
    medium · Anthropic: within 3 attempts (c@3) on each suite with 95% intervals, against the median of the API board's models, each at its highest-scoring configuration.
    • 95% interval
    • API board median
    Apex
    90%, 95% interval 70–97%
    Salesforce APIs
    89%, 95% interval 67–97%
    CI/CD
    94%, 95% interval 73–99%
    Salesforce CLI
    85%, 95% interval 64–95%
    Salesforce docs
    29%, 95% interval 13–53%
    fflib Enterprise Patterns
    61%, 95% interval 39–80%
    Flow
    78%, 95% interval 55–91%
    Governor limits & pushback
    95%, 95% interval 75–99%
    Lightning Web Components
    94%, 95% interval 74–99%
    Nonprofit Success Pack
    83%, 95% interval 61–94%
    Packaging
    67%, 95% interval 44–84%
    Permissions & access
    87%, 95% interval 62–96%
    Scratch org definitions
    50%, 95% interval 29–71%
    SOQL
    95%, 95% interval 76–99%
    Trigger Actions Framework
    50%, 95% interval 29–71%
      Show the numbers
      Per-suite scores of Claude Haiku 5.5's highest-scoring configuration on the API board, with the board's median
      Suitec@3 95% CIAPI medianTasks
      Apex
      90%95% CI 70–97
      		90%20
      Salesforce APIs
      89%95% CI 67–97
      		72%18
      CI/CD
      94%95% CI 73–99
      		88%17
      Salesforce CLI
      85%95% CI 64–95
      		90%20
      Salesforce docs
      29%95% CI 13–53
      		29%17
      fflib Enterprise Patterns
      61%95% CI 39–80
      		61%18
      Flow
      78%95% CI 55–91
      		69%18
      Governor limits & pushback
      95%95% CI 75–99
      		89%19
      Lightning Web Components
      94%95% CI 74–99
      		100%18
      Nonprofit Success Pack
      83%95% CI 61–94
      		81%18
      Packaging
      67%95% CI 44–84
      		69%18
      Permissions & access
      87%95% CI 62–96
      		87%15
      Scratch org definitions
      50%95% CI 29–71
      		50%18
      SOQL
      95%95% CI 76–99
      		100%20
      Trigger Actions Framework
      50%95% CI 29–71
      		86%18

      By reasoning effort

      The same configuration at several efforts, each against its lowest reasoning effort on the same tasks (a paired 95% interval), and charted against the effort tier and against the output tokens it costs. These scores are pass@1, the first answer alone: a paired comparison needs both efforts' answers to the same tasks, and not every effort has had its retries. Lite-subset sweeps (4 tasks per suite) are never plotted with the full benchmark. The effort study has every model.

      Claude Haiku 5.5 by reasoning effort, full benchmark: each effort against the sweep's lowest reasoning effort
      Configuration · effort · full benchmarkpass@1 95% CIAgainst the lowest reasoning effort same tasks, pairedTokens
      lowbaseline
      59%95% CI 54–64
      		—2.2k
      medium
      64%95% CI 59–69
      		clearly higher (+5 pts, 95% CI +1 pt to +9 pts)2.8k
      high
      63%95% CI 58–68
      		no clear difference (+4 pts, 95% CI −0.3 pts to +7 pts); this data can't rule out a change of up to 7 points4.5k
      Score by effort tier · full benchmark
      Overall pass@1; vertical bars are 95% confidence intervals.
      Claude Haiku 5.5, full benchmark: pass@1 by reasoning effort tierClaude Haiku 5.5: low 59%, medium 64%, high 63%0%25%50%75%100%LowMediumHighReasoning effort tierpass@1Claude Haiku 5.5
      Show the numbers
      ModelEffort (tier)pass@195% CITasksNo answerTokens: meanmedianLatency: median per answer
      Claude Haiku 5.5low59%54–642720%2.2k1.4k6.6 s
      Claude Haiku 5.5medium64%59–692720%2.8k1.7k7.7 s
      Claude Haiku 5.5high63%58–68272<1%4.5k2.5k11 s
      Score against output tokens · full benchmark
      Each point is one effort setting. Further right means more output tokens per task. The numbers are under the chart above.
      Claude Haiku 5.5, full benchmark: pass@1 against mean output tokens per taskClaude Haiku 5.5: 2.2k tokens 59%, 2.8k tokens 64%, 4.5k tokens 63%0%25%50%75%100%2k5kMean output tokens per task (log scale)pass@1lowmediumhighClaude Haiku 5.5

      In the studies

      Private evaluations

      Want this for your team?

      Azul Labs runs Forcebench privately for two kinds of Salesforce team.

      Choosing a model to run yourself?

      A private benchmark of your shortlist

      Open-weight models at the quantisations and reasoning efforts you could run on infrastructure you control, tested on your kind of Salesforce work, with a written recommendation.

      • Private tasks written from your work only ever run on local models; they are never sent to an API.
      • Re-testing after any change of model, quantisation or engine is available.
      • Typical turnaround is 2–3 weeks; large models take about a day of machine time per configuration.
      Get a private benchmarkHow it works

      Building or buying a Salesforce AI tool?

      A private evaluation of your tool

      Your tool on held-out tasks it has never seen, against API baselines (models reached through a hosted service) and Local ones (served on Forcebench’s own hardware), with a written report on where it holds up and where it fails.

      • It can start as a smaller pilot, about 30 held-out tasks with your tool against two baselines, before a full evaluation.
      • Repeat evaluations after your releases are available.
      • Typical turnaround is 2–3 weeks.
      Ask about an evaluationHow it works

      Fixed fee, quoted up front. The quote depends on the number of tasks and the number of models, tools or configurations tested.

      Or email leo@azl.au · azl.au