Suites / lwc

Lightning Web Components

Lightning Web Components engineering: public properties and reactivity, templates and conditional rendering, events, parent/child APIs, wire adapters, imperative Apex, toasts, navigation, Lightning Message Service, lightning-datatable, forms, lifecycle cleanup, slots, labels, App Builder metadata and accessibility. Each answer runs under hidden Jest tests (@salesforce/sfdx-lwc-jest); no org is needed.

18
tasks
Execution
grading
100%
highest c@3 · Claude Fable 5.1, joint first with 39 others
78
configurations with results

How it's graded

Each component runs under hidden Jest tests (@salesforce/sfdx-lwc-jest) that render it and check its behaviour. No org is needed.

Difficulty mix

5 easy8 medium5 hard

Leaderboard

Who's best at Lightning Web Components

Within three attempts (c@3) on this suite only, with 95% confidence intervals. With 18 tasks the intervals are wide: where two overlap, those configurations are not clearly different. 1 finished configuration isn’t shown yet: its retries haven’t finished. Open in the full leaderboard →

API: the best of each lab at Lightning Web Components
Within three attempts (c@3) on Lightning Web Components, on the API board (models reached through a hosted service). Bars whose intervals overlap are not clearly different.
  • 95% interval
=1Gemini 4 Argon low tie
100%, 95% interval 82–100%
=1Claude Sonnet 5 low tie
100%, 95% interval 82–100%
=1GLM-5.3 low tie
100%, 95% interval 82–100%
=1GPT-6 Astra medium tie
100%, 95% interval 82–100%
=1Muse Wonder low tie
100%, 95% interval 82–100%
=1Inkling Small high tie
100%, 95% interval 82–100%
=1Grok 4.6 high tie
100%, 95% interval 82–100%
=1DeepSeek V4.1 Flash high tie
100%, 95% interval 82–100%
=9Nemotron 3 Ultra thinking on tie
94%, 95% interval 74–99%
=9MiMo V2.6 Flash thinking on tie
94%, 95% interval 74–99%
=9Mistral Large 4 high tie
94%, 95% interval 74–99%
12Qwen3.8 Flash-Next thinking on
78%, 95% interval 55–91%
13Ling 3.0 Flash thinking on
72%, 95% interval 49–88%
14Laguna S 2.1 thinking on
67%, 95% interval 44–84%
    Show the numbers
    ModelScore95% CI
    Gemini 4 Argon low100%82–100
    Claude Sonnet 5 low100%82–100
    GLM-5.3 low100%82–100
    GPT-6 Astra medium100%82–100
    Muse Wonder low100%82–100
    Inkling Small high100%82–100
    Grok 4.6 high100%82–100
    DeepSeek V4.1 Flash high100%82–100
    Nemotron 3 Ultra thinking on94%74–99
    MiMo V2.6 Flash thinking on94%74–99
    Mistral Large 4 high94%74–99
    Qwen3.8 Flash-Next thinking on78%55–91
    Ling 3.0 Flash thinking on72%49–88
    Laguna S 2.1 thinking on67%44–84
    Local: the best of each lab at Lightning Web Components
    Within three attempts (c@3) on Lightning Web Components, on the Local board (models served on Forcebench’s own hardware). Bars whose intervals overlap are not clearly different.
    • 95% interval
    =1GLM-5.3 Flash low · EXL3 4.0bpw experts, FP8 dense tie
    100%, 95% interval 82–100%
    =1Qwen3.8 Flash-Next medium · NVFP4 tie
    100%, 95% interval 82–100%
    =1DeepSeek V4.1 Flash 1017 high · Native FP8, FP4 experts tie
    100%, 95% interval 82–100%
    4Gemma 4 31B thinking on · QAT W4A16
    89%, 95% interval 67–97%
    =5Muse Glimmer 30B high · NVFP4 + FP8 mixed (NVIDIA ModelOpt) tie
    83%, 95% interval 61–94%
    =5MiniMax M3.1 Flash high · NVFP4 experts, MXFP8 (QAT) tie
    83%, 95% interval 61–94%
    =5Nemotron 3.5 Super VL thinking on · NVFP4 mixed precision (NVIDIA ModelOpt) tie
    83%, 95% interval 61–94%
    8MiMo V2.6 Flash thinking on · FP8 + MXFP4 experts
    78%, 95% interval 55–91%
    9Ornith 1.5 122B thinking on · NVFP4 experts, FP8 attention (NVIDIA ModelOpt)
    72%, 95% interval 49–88%
    10MiniCPM-V 4.7 35B-A3B thinking on · INT4 W4A16 experts, BF16 dense
    50%, 95% interval 29–71%
      Show the numbers
      ModelScore95% CI
      GLM-5.3 Flash low · EXL3 4.0bpw experts, FP8 dense100%82–100
      Qwen3.8 Flash-Next medium · NVFP4100%82–100
      DeepSeek V4.1 Flash 1017 high · Native FP8, FP4 experts100%82–100
      Gemma 4 31B thinking on · QAT W4A1689%67–97
      Muse Glimmer 30B high · NVFP4 + FP8 mixed (NVIDIA ModelOpt)83%61–94
      MiniMax M3.1 Flash high · NVFP4 experts, MXFP8 (QAT)83%61–94
      Nemotron 3.5 Super VL thinking on · NVFP4 mixed precision (NVIDIA ModelOpt)83%61–94
      MiMo V2.6 Flash thinking on · FP8 + MXFP4 experts78%55–91
      Ornith 1.5 122B thinking on · NVFP4 experts, FP8 attention (NVIDIA ModelOpt)72%49–88
      MiniCPM-V 4.7 35B-A3B thinking on · INT4 W4A16 experts, BF16 dense50%29–71
      Lightning Web Components: c@3 by configuration with 95% confidence intervals
      #Model and configScore c@3 · 95% CITasks
      1Claude Fable 5.1 effort high
      100%95% CI 82–100
      		18
      1Claude Haiku 5.5 effort high
      100%95% CI 82–100
      		18
      1Claude Haiku 5.5 effort low
      100%95% CI 82–100
      		18
      1Claude Opus 4.8 effort high
      100%95% CI 82–100
      		18
      1Claude Opus 4.8 effort low
      100%95% CI 82–100
      		18
      1Claude Opus 4.8 thinking off
      100%95% CI 82–100
      		18
      1Claude Opus 5.5 effort high
      100%95% CI 82–100
      		18
      1Claude Opus 5.5 effort medium
      100%95% CI 82–100
      		18
      1Claude Opus 5.5 effort low
      100%95% CI 82–100
      		18
      1Claude Sonnet 5 effort high
      100%95% CI 82–100
      		18
      1Claude Sonnet 5 effort medium
      100%95% CI 82–100
      		18
      1Claude Sonnet 5 thinking off
      100%95% CI 82–100
      		18
      1Claude Sonnet 5 effort low
      100%95% CI 82–100
      		18
      1Claude Sonnet 5.5 effort high
      100%95% CI 82–100
      		18
      1Claude Sonnet 5.5 effort low
      100%95% CI 82–100
      		18
      1Claude Sonnet 5.5 effort medium
      100%95% CI 82–100
      		18
      1DeepSeek V4 Pro FP8 · effort high
      100%95% CI 82–100
      		18
      1DeepSeek V4.1 Flash FP8 · effort high
      100%95% CI 82–100
      		18
      1DeepSeek V4.1 Flash Native FP8, FP4 experts · SGLang · effort high
      100%95% CI 82–100
      		18
      1DeepSeek V4.1 Flash 1017 Native FP8, FP4 experts · SGLang · effort high
      100%95% CI 82–100
      		18
      1Gemini 3.8 Flash effort medium
      100%95% CI 82–100
      		18
      1Gemini 4 Argon effort medium
      100%95% CI 82–100
      		18
      1Gemini 4 Argon effort low
      100%95% CI 82–100
      		18
      1Gemini 4 Argon effort high
      100%95% CI 82–100
      		18
      1GLM-5.3 FP8 · effort low
      100%95% CI 82–100
      		18
      1GLM-5.3 Flash EXL3 4.0bpw experts, 4-bit dense · TensorFold · effort high
      100%95% CI 82–100
      		18
      1GLM-5.3 Flash EXL3 4.0bpw experts, FP8 dense · vLLM + ExLlamaV3 · effort low
      100%95% CI 82–100
      		18
      1GPT-5.6 Sol effort medium
      100%95% CI 82–100
      		18
      1GPT-6 Astra effort medium
      100%95% CI 82–100
      		18
      1GPT-6 Astra effort high
      100%95% CI 82–100
      		18
      1GPT-6 Luna effort medium
      100%95% CI 82–100
      		18
      1GPT-6.1 Sol effort high
      100%95% CI 82–100
      		18
      1Grok 4.6 effort high
      100%95% CI 82–100
      		18
      1Grok 4.7 effort high
      100%95% CI 82–100
      		18
      1Inkling Small FP8 · effort high
      100%95% CI 82–100
      		18
      1Muse Wonder effort medium
      100%95% CI 82–100
      		18
      1Muse Wonder effort max
      100%95% CI 82–100
      		18
      1Muse Wonder effort low
      100%95% CI 82–100
      		18
      1Muse Wonder effort high
      100%95% CI 82–100
      		18
      1Qwen3.8 Flash-Next NVFP4 · vLLM · effort medium
      100%95% CI 82–100
      		18
      41Claude Haiku 5.5 effort medium
      94%95% CI 74–99
      		18
      41Claude Opus 4.8 effort medium
      94%95% CI 74–99
      		18
      41DeepSeek V4 Flash Vision experimental FP8 + FP4 experts · vLLM · effort high
      94%95% CI 74–99
      		18
      41GLM-5.3 Flash FP8 · effort high
      94%95% CI 74–99
      		18
      41GLM-5.3 Flash EXL3 4.0bpw experts, FP8 dense · vLLM + ExLlamaV3 · effort high
      94%95% CI 74–99
      		18
      41MiMo V2.6 Flash FP8 · thinking on
      94%95% CI 74–99
      		18
      41MiMo V2.6 Pro FP8 · thinking on
      94%95% CI 74–99
      		18
      41Mistral Large 4 effort high
      94%95% CI 74–99
      		18
      41Nemotron 3 Ultra thinking on
      94%95% CI 74–99
      		18
      41Qwen3.8 Flash-Next INT4-AutoRound, 5 of 10 experts · TensorFold (Zig) · effort medium
      94%95% CI 74–99
      		18
      51Claude Haiku 4.5 thinking on
      89%95% CI 67–97
      		18
      51DeepSeek V4.1 Flash EXL3 2.9bpw · vLLM + ExLlamaV3 · effort high
      89%95% CI 67–97
      		18
      51Gemma 4 31B QAT W4A16 · vLLM · thinking on
      89%95% CI 67–97
      		18
      51Qwen3.8 Flash-Next MLX 4-bit · MTPLX · effort medium
      89%95% CI 67–97
      		18
      51Qwen3.8 Flash-Next MLX 4-bit · TensorFold · effort medium
      89%95% CI 67–97
      		18
      51Qwen3.8 Flash-Next INT4-AutoRound, 5 of 10 experts · TensorFold (Zig) · effort medium
      89%95% CI 67–97
      		18
      57Gemma 4 26B-A4B NVFP4 · vLLM · thinking on
      83%95% CI 61–94
      		18
      57MiniMax M3.1 Flash NVFP4 experts, MXFP8 (QAT) · vLLM · effort high
      83%95% CI 61–94
      		18
      57Muse Glimmer 30B NVFP4 + FP8 mixed (NVIDIA ModelOpt) · vLLM · effort high
      83%95% CI 61–94
      		18
      57Nemotron 3.5 Super VL NVFP4 mixed precision (NVIDIA ModelOpt) · TensorFold (Zig) · thinking on
      83%95% CI 61–94
      		18
      61Claude Haiku 4.5 thinking off
      78%95% CI 55–91
      		18
      61MiMo V2.6 Flash FP8 + MXFP4 experts · SGLang · thinking on
      78%95% CI 55–91
      		18
      61Qwen3.8 Flash-Next thinking on
      78%95% CI 55–91
      		18
      64GLM-5.3 Flash EXL3 4.0bpw experts, FP8 dense · vLLM + ExLlamaV3 · thinking off
      72%95% CI 49–88
      		18
      64Ling 3.0 Flash BF16 · thinking on
      72%95% CI 49–88
      		18
      64Ornith 1.5 122B NVFP4 experts, FP8 attention (NVIDIA ModelOpt) · vLLM · thinking on
      72%95% CI 49–88
      		18
      67Laguna S 2.1 FP4 · thinking on
      67%95% CI 44–84
      		18
      68Qwen3.8 27B AWQ-INT4 · vLLM · effort low
      61%95% CI 39–80
      		18
      69MiniCPM-V 4.7 35B-A3B INT4 W4A16 experts, BF16 dense · vLLM · thinking on
      50%95% CI 29–71
      		18
      69Qwen3.8 27B MLX 4-bit · MTPLX · effort low
      50%95% CI 29–71
      		18
      71MiniCPM-V 4.7 35B-A3B BF16 · vLLM · thinking on
      44%95% CI 25–66
      		18
      71Qwen3.6 35B-A3B NVFP4 · vLLM · thinking on
      44%95% CI 25–66
      		18
      71Qwen3.8 27B Splash 4-bit · Splash · effort low
      44%95% CI 25–66
      		18
      74Qwen3.8 27B MLX 8-bit · MTPLX · effort low
      39%95% CI 20–61
      		18
      74Qwen3.8 27B AWQ-INT4 · vLLM · effort xhigh
      39%95% CI 20–61
      		18
      76Qwen3.8 27B AWQ-INT4 · vLLM · effort medium
      33%95% CI 16–56
      		18
      77Nemotron 3.5 Lightning NVFP4 · thinking on
      11%95% CI 3–33
      		18
      78Nemotron 3.5 Bolt IQ2_XXS + NVFP4 + Q8_0 mixed (NVIDIA ModelOpt) · TensorFold (Zig) · thinking on
      6%95% CI 1–26
      		18

      Tasks

      What's in the suite

      Solve rate is the share of finished runs whose first answer to the task passed (pass@1), a rough guide to how hard models find it. Difficulty is the task author’s label; from results is what the solve rate says: easy when two thirds or more of the finished runs pass it, hard at a third or fewer, once five have graded it.

      Lightning Web Components tasks with difficulty and solve rate
      TaskDifficultyFrom resultsSolve rate
      Accessible collapsible section with slotslwc-collapsible-section-slotseasyeasy
      72%
      Child-to-parent select event from a product tilelwc-product-tile-select-eventeasyeasy
      82%
      Reactive case status badgelwc-case-status-badgeeasyeasy
      76%
      Sorted contact list from a public arraylwc-contact-list-sortedeasyeasy
      75%
      Translatable, accessible favourite togglelwc-favorite-toggle-labelseasymedium
      62%
      Account header with the getRecord wire adapterlwc-account-header-wiremediummedium
      62%
      Debounced search with stale-response protectionlwc-contact-search-debouncemediumeasy
      84%
      Expose a component to Lightning App Builderlwc-pipeline-card-app-buildermediummedium
      56%
      Form validation with custom validity and toastslwc-registration-form-validationmediumeasy
      80%
      Imperative Apex with loading, empty and error stateslwc-open-opportunities-imperative-apexmediumeasy
      77%
      Lightning Message Service subscriber with clean lifecyclelwc-lms-account-panelmediummedium
      46%
      NavigationMixin page references and generated URLslwc-account-quick-links-navigationmediummedium
      52%
      Parent calling child @api methodslwc-order-lines-child-apimediumeasy
      75%
      Datatable row actions with confirm, delete, navigation and toastslwc-opportunity-row-actionshardmedium
      66%
      Fix reactivity and data-flow bugs in a checklist componentlwc-task-checklist-reactivity-fixhardeasy
      73%
      Inline-edit datatable saved through LDS with refreshApexlwc-contacts-datatable-inline-savehardeasy
      68%
      Keyboard-accessible listbox with aria-activedescendantlwc-status-listbox-keyboardhardmedium
      57%
      Polling component with visibility pause and lifecycle cleanuplwc-queue-monitor-polling-lifecyclehardeasy
      80%

      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