LLM model picker
Describe the job and get a shortlist of models, with the reasons behind each score and the trade-off to the runner-up.
Free, runs in your browser, no sign-up. Updated 2026-10-25.
Guidance, not a benchmark. Picks come from published positioning, prices and stated rules, not quality tests. Try the shortlist on your own prompts.
Result
Suggested: deepseek-v4-pro (DeepSeek), runner-up MiniMax M3. MiniMax M3 is cheaper ($20.58 vs $71.19 a month), and it is the same balanced class. Test both on your own prompts before choosing. Guidance from stated rules and published prices, not a benchmark.
Runner-up: MiniMax M3. MiniMax M3 is cheaper ($20.58 vs $71.19 a month), and it is the same balanced class. Test both on your own prompts before choosing.
Why deepseek-v4-pro
- Classed balanced: Named pro; mid price band. Half price off-peak.
- Customer chat and support: balanced models get 3 for fit. Needs good judgement at conversational speed and volume, so the mid class is favoured.
- Interactive (chat, a few seconds is fine): +0.5. Small preference for quicker classes.
- About $71.19 a month for 100,000,000 tokens at this mix (price checked 2026-10-23).
| Model | Score | Fit | Speed | Budget | Cost | Penalty | Per month |
|---|---|---|---|---|---|---|---|
| deepseek-v4-proDeepSeek · balanced | 4.23 | 3 | 0.5 | 0 | 0.73 | 0 | $71.19 |
| MiniMax M3MiniMax · balanced | 4.14 | 3 | 0.5 | 0 | 1.14 | −0.5 | $20.58 |
| Gemini 2.5 ProGoogle · balanced | 4.11 | 3 | 0.5 | 0 | 0.61 | 0 | $104 |
| Claude Sonnet 5.5Anthropic · balanced | 4.04 | 3 | 0.5 | 0 | 0.54 | 0 | $130 |
| gpt-6.1-solOpenAI · balanced | 4.04 | 3 | 0.5 | 0 | 0.54 | 0 | $130 |
| Claude Sonnet 5Anthropic · balanced | 4.02 | 3 | 0.5 | 0 | 0.52 | 0 | $136 |
| gpt-5.6-terraOpenAI · balanced | 4 | 3 | 0.5 | 0 | 0.5 | 0 | $146 |
| Gemini 3.1 Pro PreviewGoogle · balanced | 4 | 3 | 0.5 | 0 | 0.5 | 0 | $146 |
| Claude Haiku 5.5Anthropic · fast | 4 | 2 | 0.5 | 0 | 1.5 | 0 | $6.78 |
| gpt-6-lunaOpenAI · fast | 4 | 2 | 0.5 | 0 | 1.5 | 0 | $6.78 |
| MiniMax M3.1-FlashMiniMax · fast | 3.89 | 2 | 0.5 | 0 | 1.39 | 0 | $9.42 |
| gpt-5.6-lunaOpenAI · fast | 3.75 | 2 | 0.5 | 0 | 1.25 | 0 | $14.56 |
| deepseek-flashDeepSeek · fast | 3.69 | 2 | 0.5 | 0 | 1.19 | 0 | $17.45 |
| Gemini 2.5 FlashGoogle · fast | 3.57 | 2 | 0.5 | 0 | 1.07 | 0 | $25.34 |
| Gemini 3.8 FlashGoogle · fast | 3.34 | 2 | 0.5 | 0 | 0.84 | 0 | $50.85 |
| gpt-5.4-miniOpenAI · fast | 3.32 | 2 | 0.5 | 0 | 0.82 | 0 | $54.60 |
| Claude Haiku 4.5Anthropic · fast | 3.25 | 2 | 0.5 | 0 | 0.75 | 0 | $67.80 |
| Gemini 4 ArgonGoogle · frontier | 1.54 | 1 | 0 | 0 | 0.54 | 0 | $130 |
| Claude Opus 5.5Anthropic · frontier | 1.31 | 1 | 0 | 0 | 0.31 | 0 | $260 |
| gpt-5.6-solOpenAI · frontier | 1.3 | 1 | 0 | 0 | 0.3 | 0 | $271 |
| Claude Opus 5Anthropic · frontier | 1.23 | 1 | 0 | 0 | 0.23 | 0 | $339 |
| Claude Fable 5.1Anthropic · frontier | 1.02 | 1 | 0 | 0 | 0.02 | 0 | $635 |
| gpt-6-astraOpenAI · frontier | 1 | 1 | 0 | 0 | 0 | 0 | $678 |
Guidance from rules and published prices, not a benchmark. Context-window limits are not checked here; confirm them on the provider's model page.
Ask OceanLab to test models on your prompts
Opens an email to OceanLab with this result and a link to your inputs. Nothing is sent until you press send, and there is no sign-up.
How this is calculated
Score = task fit + latency bonus + budget bonus + cost weight × cost score − penalties. The cost score runs from 0 to 1 on a log scale between the dearest and cheapest model for your monthly volume.
Tiers come from each provider's own positioning of the model or, where none was quoted, from its name and price band. The weights are OceanLab planning assumptions, not benchmark results.
Penalties apply when your prompts exceed a size where the model's price goes up, or when a price is flagged for re-checking. Context-window limits are not checked.
Sources
- Anthropic API pricing · checked 2026-10-23
- OpenAI API pricing · checked 2026-10-23
- Google API pricing · checked 2026-10-23
- DeepSeek API pricing · checked 2026-10-23
- MiniMax API pricing · checked 2026-10-23
Questions about this tool
Is this based on benchmarks?
No. It uses published rules: the provider's own positioning of each model, its price for your workload, and long-context surcharges from the pricing pages. Test the shortlist on your own prompts.
Why show a runner-up?
The top two are often close. The trade-off line says what you give up or gain by picking the second.
Can I see the scoring?
Yes. Each model shows its points for task fit, latency, budget, cost and any penalties, with the reasons.
Is anything I enter sent anywhere?
No. The tool runs in your browser and there is no sign-up. Nothing leaves the page unless you choose to email your result to OceanLab.
Other free tools
OceanLab product builds like this
Products OceanLab designed and built. Each case study covers the screens, the scope and the rules it was built for.
Sonde - LLM Evaluation and Observability Platform with Release Gates
Designed and built a clickable LLM evaluation and observability platform for the AI team of a French health mutual: redacted production traces span by span, release gates that block a prompt or model change when quality regresses, alerts with owners and ageing, and the monitoring record its risk and compliance teams can show a supervisor.
Tekmė - No-Code Workflow Automation Builder with AI Steps
Designed and built a clickable no-code automation builder for a Vilnius SaaS startup selling to EU operations teams: a node canvas with branches, error paths and AI steps that classify, extract, summarise or act as an agent, a cost estimate per run, and a run history where every step is recorded with personal data masked.
Lektor - AI Code Review Agent That Comments but Never Merges, Hosted Inside a Private Bank
Designed and built a clickable AI code review agent for the engineering team of a Zurich private bank: it reviews every merge request on GitLab and GitHub Enterprise with models hosted in the bank's own tenancy, masks client data and secrets before any model sees the diff, posts inline findings with rule, CWE and suggested patch, and never approves or merges.
Kaganek - RAG Knowledge Assistant with Citations and Permissions Intact
Designed and built a clickable internal knowledge assistant for a Kraków consulting engineering firm: staff ask in Polish or English and get answers built only from the firm's own SharePoint, Confluence, Google Drive and file share documents, with every claim linked to its source passage and an honest "I don't know" when the sources are weak.
