Suites / scratch-def
Scratch org definitions
Turn a client brief into a working scratch org definition file (and sfdx-project.json). Answers are validated against the official JSON schemas, the documented scratch org feature list and the Metadata API settings types; settings are check-only deployed to a dedicated grader org exactly as scratch org creation applies them (types with lasting side effects, such as fiscal year or currencies, are validated offline). Task rules then check the brief's requirements (edition, features, settings values, locale).
How it's graded
Definition files are validated against the official JSON schemas and the documented scratch org feature list, and their settings are deployed check-only to a grader org exactly as scratch org creation applies them. Task rules then check the brief: edition, features, settings values and locale.
Difficulty mix
Leaderboard
Who's best at Scratch org definitions
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 →
| # | Model and config | Score c@3 · 95% CI | Tasks |
|---|---|---|---|
| 1 | Claude Opus 5.5 effort high | 94%95% CI 74–99 | 18 |
| 1 | Claude Opus 5.5 effort low | 94%95% CI 74–99 | 18 |
| 3 | Claude Fable 5.1 effort high | 83%95% CI 61–94 | 18 |
| 3 | Claude Opus 5.5 effort medium | 83%95% CI 61–94 | 18 |
| 3 | Grok 4.7 effort high | 83%95% CI 61–94 | 18 |
| 6 | Claude Sonnet 5.5 effort high | 78%95% CI 55–91 | 18 |
| 6 | GPT-5.6 Sol effort medium | 78%95% CI 55–91 | 18 |
| 8 | Claude Sonnet 5.5 effort low | 67%95% CI 44–84 | 18 |
| 8 | GPT-6 Astra effort medium | 67%95% CI 44–84 | 18 |
| 10 | DeepSeek V4.1 Flash FP8 · effort high | 61%95% CI 39–80 | 18 |
| 10 | Gemini 3.8 Flash effort medium | 61%95% CI 39–80 | 18 |
| 10 | GPT-6.1 Sol effort high | 61%95% CI 39–80 | 18 |
| 10 | Grok 4.6 effort high | 61%95% CI 39–80 | 18 |
| 14 | Claude Opus 4.8 effort high | 56%95% CI 34–75 | 18 |
| 14 | Claude Sonnet 5.5 effort medium | 56%95% CI 34–75 | 18 |
| 14 | GPT-6 Luna effort medium | 56%95% CI 34–75 | 18 |
| 17 | Claude Haiku 5.5 effort medium | 50%95% CI 29–71 | 18 |
| 17 | Claude Opus 4.8 effort medium | 50%95% CI 29–71 | 18 |
| 17 | Claude Sonnet 5 effort high | 50%95% CI 29–71 | 18 |
| 17 | Claude Sonnet 5 effort medium | 50%95% CI 29–71 | 18 |
| 17 | GPT-6 Astra effort high | 50%95% CI 29–71 | 18 |
| 22 | Claude Opus 4.8 effort low | 44%95% CI 25–66 | 18 |
| 22 | DeepSeek V4.1 Flash EXL3 2.9bpw · vLLM + ExLlamaV3 · effort high | 44%95% CI 25–66 | 18 |
| 22 | GLM-5.3 Flash FP8 · effort high | 44%95% CI 25–66 | 18 |
| 25 | Claude Haiku 5.5 effort high | 39%95% CI 20–61 | 18 |
| 25 | Claude Haiku 5.5 effort low | 39%95% CI 20–61 | 18 |
| 25 | Claude Opus 4.8 thinking off | 39%95% CI 20–61 | 18 |
| 25 | DeepSeek V4 Flash Vision experimental FP8 + FP4 experts · vLLM · effort high | 39%95% CI 20–61 | 18 |
| 25 | DeepSeek V4 Pro FP8 · effort high | 39%95% CI 20–61 | 18 |
| 25 | GLM-5.3 Flash EXL3 4.0bpw experts, FP8 dense · vLLM + ExLlamaV3 · effort high | 39%95% CI 20–61 | 18 |
| 25 | MiMo V2.6 Pro FP8 · thinking on | 39%95% CI 20–61 | 18 |
| 25 | Mistral Large 4 effort high | 39%95% CI 20–61 | 18 |
| 33 | Claude Sonnet 5 thinking off | 33%95% CI 16–56 | 18 |
| 33 | Inkling Small FP8 · effort high | 33%95% CI 16–56 | 18 |
| 35 | Claude Sonnet 5 effort low | 28%95% CI 13–51 | 18 |
| 35 | Nemotron 3 Ultra thinking on | 28%95% CI 13–51 | 18 |
| 37 | GLM-5.3 Flash EXL3 4.0bpw experts, FP8 dense · vLLM + ExLlamaV3 · effort low | 22%95% CI 9–45 | 18 |
| 37 | GLM-5.3 Flash EXL3 4.0bpw experts, FP8 dense · vLLM + ExLlamaV3 · thinking off | 22%95% CI 9–45 | 18 |
| 39 | DeepSeek V4.1 Flash Native FP8, FP4 experts · SGLang · effort high | 17%95% CI 6–39 | 18 |
| 39 | Qwen3.6 35B-A3B NVFP4 · vLLM · thinking on | 17%95% CI 6–39 | 18 |
| 39 | Qwen3.8 Flash-Next NVFP4 · vLLM · effort medium | 17%95% CI 6–39 | 18 |
| 42 | Claude Haiku 4.5 thinking on | 11%95% CI 3–33 | 18 |
| 42 | Claude Haiku 4.5 thinking off | 11%95% CI 3–33 | 18 |
| 42 | Gemma 4 26B-A4B NVFP4 · vLLM · thinking on | 11%95% CI 3–33 | 18 |
| 42 | Ling 3.0 Flash BF16 · thinking on | 11%95% CI 3–33 | 18 |
| 42 | MiMo V2.6 Flash FP8 + MXFP4 experts · SGLang · thinking on | 11%95% CI 3–33 | 18 |
| 42 | MiMo V2.6 Flash FP8 · thinking on | 11%95% CI 3–33 | 18 |
| 42 | Qwen3.8 27B MLX 4-bit · MTPLX · effort low | 11%95% CI 3–33 | 18 |
| 42 | Qwen3.8 27B AWQ-INT4 · vLLM · effort low | 11%95% CI 3–33 | 18 |
| 42 | Qwen3.8 27B AWQ-INT4 · vLLM · effort xhigh | 11%95% CI 3–33 | 18 |
| 42 | Qwen3.8 Flash-Next INT4-AutoRound, 5 of 10 experts · TensorFold (Zig) · effort medium | 11%95% CI 3–33 | 18 |
| 42 | Qwen3.8 Flash-Next thinking on | 11%95% CI 3–33 | 18 |
| 53 | Gemma 4 31B QAT W4A16 · vLLM · thinking on | 6%95% CI 1–26 | 18 |
| 53 | GLM-5.3 Flash EXL3 4.0bpw experts, 4-bit dense · TensorFold · effort high | 6%95% CI 1–26 | 18 |
| 53 | Nemotron 3.5 Lightning NVFP4 · thinking on | 6%95% CI 1–26 | 18 |
| 53 | Qwen3.8 27B MLX 8-bit · MTPLX · effort low | 6%95% CI 1–26 | 18 |
| 53 | Qwen3.8 27B AWQ-INT4 · vLLM · effort medium | 6%95% CI 1–26 | 18 |
| 53 | Qwen3.8 27B Splash 4-bit · Splash · effort low | 6%95% CI 1–26 | 18 |
| 53 | Qwen3.8 Flash-Next MLX 4-bit · MTPLX · effort medium | 6%95% CI 1–26 | 18 |
| 53 | Qwen3.8 Flash-Next MLX 4-bit · TensorFold · effort medium | 6%95% CI 1–26 | 18 |
| 61 | Laguna S 2.1 FP4 · thinking on | 0%95% CI 0–18 | 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.
| Task | Difficulty | From results | Solve rate |
|---|---|---|---|
| B2C insurer with person accountsscratch-def-person-accounts | easy | easy | 71% |
| Integration org with state and country picklistsscratch-def-state-country-picklists | easy | hard | 31% |
| ISV regression org on the preview releasescratch-def-partner-preview-release | easy | hard | 18% |
| Professional Edition org with API and Apexscratch-def-professional-edition-api | easy | hard | 16% |
| sfdx-project.json for a sandbox-based teamscratch-def-sfdx-project-sandbox-team | easy | easy | 68% |
| Australian locale and financial yearscratch-def-australian-locale | medium | hard | 5% |
| Experience Cloud org for a guest-accessible sitescratch-def-experience-cloud-guest-site | medium | hard | 6% |
| German subsidiary with multi-currencyscratch-def-multicurrency-germany | medium | hard | 2% |
| Migrate a definition off orgPreferencesscratch-def-fix-org-preferences | medium | hard | 26% |
| Scratch org for Agentforce and Prompt Builder workscratch-def-einstein-generative-ai | medium | hard | 16% |
| Service Cloud console with Lightning Knowledge and Omni-Channelscratch-def-service-knowledge-omni | medium | hard | 10% |
| Team selling with account and opportunity teamsscratch-def-opportunity-teams | medium | hard | 26% |
| Field Service org with licences and multi-currencyscratch-def-field-service-multicurrency | hard | hard | 0% |
| Repair a scratch org definition that fails creationscratch-def-fix-broken-definition | hard | hard | 0% |
| Sales Cloud forecasting and pipeline orgscratch-def-sales-forecasting | hard | hard | 0% |
| sfdx-project.json for a two-package 2GP monoreposcratch-def-sfdx-project-2gp-monorepo | hard | hard | 26% |
| Sharing and record type prerequisites for a package installscratch-def-package-object-settings | hard | hard | 13% |
| Shield and hardened security for a health clientscratch-def-shield-security | hard | hard | 8% |
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.
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.
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