Rank AI models for a workload using task fit, context, budget, benchmark coverage, and provider availability.
Interactive Tool
| Result | Value | How to read it |
|---|---|---|
| Eligible models | 200 | Models that pass the selected evidence and constraint gates. |
| Ranked shortlist | 20 | Top candidates shown below. |
| Priced candidates | 136 | Models with both input and output price minima. |
| Rank | Model | Fit score | Context | Input / 1M | Workload cost | Published benchmark rank | Providers |
|---|---|---|---|---|---|---|---|
| 1 | Gemini 3.1 ProGoogle DeepMind | 85.44 | 1.05M tokens | $2.00 / 1M | $440.00 | 3 | 2 |
| 2 | GPT-6.1 SolOpenAI | 83.33 | 1.05M tokens | $2.00 / 1M | $400.00 | 4 | 2 |
| 3 | GPT-5.6 TerraOpenAI | 82.28 | 1.05M tokens | $2.00 / 1M | $440.00 | 5 | 2 |
| 4 | Claude Sonnet 5Anthropic | 81.67 | 1M tokens | $2.00 / 1M | $400.00 | 6 | 3 |
| 5 | GPT-6 AstraOpenAI | 80.43 | 1.05M tokens | $10.00 / 1M | $2,000.00 | 2 | 2 |
| 6 | Gemini 3.8 FlashGoogle DeepMind | 80.21 | 1.05M tokens | $0.75 / 1M | $150.00 | 7 | 2 |
| 7 | Claude Fable 5.1Anthropic | 79.83 | 1M tokens | $10.00 / 1M | $2,000.00 | 1 | 1 |
| 8 | GPT-6 LunaOpenAI | 79.81 | 1.05M tokens | $0.10 / 1M | $20.00 | 8 | 2 |
| 9 | InklingThinking Machines Lab | 79.52 | 1.05M tokens | $0.95 / 1M | $176.00 | 11 | 4 |
| 10 | MiniMax M3MiniMax | 79.23 | 1.05M tokens | $0.28 / 1M | $50.00 | 12 | 5 |
| 11 | Qwen3.8 MaxQwen | 78.72 | 1M tokens | $1.65 / 1M | $264.02 | 10 | 4 |
| 12 | Kimi K3Moonshot AI | 76.62 | 1.05M tokens | $0.24 / 1M | $262.62 | 14 | 5 |
| 13 | Claude Opus 5.5Anthropic | 74.59 | 1M tokens | $4.00 / 1M | $800.00 | 9 | 2 |
| 14 | Qwen3.8 27BQwen | 72.58 | 262.14K tokens | $0.20 / 1M | $64.00 | 16 | 3 |
| 15 | Grok 4.20 Multi AgentxAI | 71.85 | 1M tokens | $1.25 / 1M | $175.00 | 13 | 1 |
| 16 | Gemini 3.5 Flash LiteGoogle DeepMind | 71.48 | 1.05M tokens | $0.30 / 1M | $80.00 | 17 | 2 |
| 17 | GLM 5.3 FlashZ.ai | 71.11 | 1M tokens | $0.08 / 1M | $12.50 | 18 | 4 |
| 18 | Grok 4.7xAI | 69.51 | 500K tokens | $2.00 / 1M | $320.00 | 15 | 1 |
| 19 | DeepSeek V4.1 FlashDeepSeek | 68.65 | 1.05M tokens | $0.15 / 1M | $27.00 | 19 | 3 |
| 20 | GPT-5.4 MiniOpenAI | 67.09 | 400K tokens | $0.75 / 1M | $165.00 | 20 | 2 |
Fit scores are transparent screening aids, not universal quality claims. Unknown capability, context, or price fields are never filled by inference.
Inputs
| Input | How it is used |
|---|---|
| Workload | General, coding, reasoning, vision, or structured-output work. |
| Context requirement | The minimum input and output capacity the request needs. |
| Monthly volume | Input and output tokens used to compare estimated spend. |
| Priority | Balance documented performance, cost, or context capacity. |
Candidates must pass the selected capability and context gates. The ranking then combines primary-source specifications, current catalog price minima, provider breadth, and published benchmark rank where available.
Treat the first rows as a research queue. Open each model record to verify the exact provider endpoint, price scope, source date, and any unknown fields before production use.
Boundaries
Catalog values retain their source and freshness on the linked model, provider, benchmark, or comparison page. Editable scenario assumptions are not Model Markets measurements.
Continue the Analysis
| Tool | Next question |
|---|---|
| Price vs Performance | Screen for models that combine useful published performance with acceptable token economics. |
| LLM Cost Calculator | Price one explicit token workload against an exact published provider endpoint. |
| Model Router | Split routine and difficult requests across eligible models without pretending one model is optimal for every call. |
Questions
Turn workload constraints into a short, inspectable model shortlist instead of a universal best-model claim. It returns a ranked shortlist with fit reasons, estimated workload cost, sourced limits, and direct model links.
Where the calculation needs model facts, it uses the current Model Markets catalog snapshot updated 2026-10-19. User-entered assumptions remain clearly editable, and unsupported values stay unknown rather than being inferred.
A low published price may belong to only one provider or tier. Benchmark coverage differs by model and missing results are not zero quality. The selector does not measure your prompt quality, reliability, or regional latency. Open the linked canonical records and primary sources before making a production or purchasing decision.
