Find your next model

Choose a task, estimate your monthly tokens, and build a shortlist with independent benchmark scores.

255Models priced
Artificial Analysis ↗

Broad text tasks, such as drafting and summarising. This selects a published index; test the shortlist on your own prompts.

Start with a workload

Optional illustrative assumptions. Edit the monthly totals to fit your usage. Each example starts with 0% cache.

Short text assistance

1,000 requests/month × (2,000 input + 500 output tokens per request)

= 2M input + 0.5M output / month

Coding assistance

500 requests/month × (20,000 input + 2,000 output tokens per request)

= 10M input + 1M output / month

Reasoning exercises

200 requests/month × (4,000 input + 2,000 output tokens per request)

= 0.8M input + 0.4M output / month

Models for your workload

255 matching · 37 with a benchmark

Monthly token estimates at MiniRouter catalog rates. Only published cache discounts apply; extra reasoning tokens and cache writes are not modelled. Per-request minimums need request-level usage.

Methodology & limitations ↓

Highest task benchmark

Claude Fable 5.2

Highest reported selected-task score among your matches.

61.8 Intelligence Index points · $168.00 / month

Context
1,000,000 tokens
Tools
Listed in catalog
Open weights
Unknown
Evaluation & evidence

Reported evaluation: Claude Fable 5.2 (High, Default Fallback) ↗

Artificial Analysis Intelligence Index · Version 5

Retrieved: . Measured: Unknown.

Score source: Source record ↗

Compatible API settings: Unknown. The reported evaluation settings cannot be mapped to a supported catalog configuration.

Token estimate only; cost per acceptable result is unavailable.

Lowest estimated cost with a task score

GPT-4.1 nano

Lowest token estimate among your scored matches; not a quality-per-dollar ranking.

7.8 Intelligence Index points · $1.89 / month

Context
1,047,576 tokens
Tools
Listed in catalog
Open weights
Unknown
Evaluation & evidence

Reported evaluation: GPT-4.1 nano ↗

Artificial Analysis Intelligence Index · Version unknown

Retrieved: . Measured: Unknown.

Score source: Source record ↗

Compatible API settings: Unknown. The reported evaluation settings cannot be mapped to a supported catalog configuration.

Token estimate only; cost per acceptable result is unavailable.

Index versions are unknown or differ across these matches; scores may not be directly comparable. Independent benchmark scores do not establish quality on your task or cost per acceptable result.

ModelEst. / month
$168.00
$42.00
$210.00
$31.50
$105.00
$210.00
$33.60
$63.00
$84.00
$33.60

Score vs price

Intelligence Index against estimated monthly token cost for your workload. Up and left is better.

404550556065$50$100$200Estimated USD / month, log scaleClaude Fable 5.2Gemini 4 CarbonGPT-6 AstraGemini 4 HydrogenClaude Opus 5Claude Fable 5.1Grok 4.6Kimi K3
  • Claude Fable 5.2: 61.8 points, $168 / month (MiniRouter token estimate)
  • Gemini 4 Carbon: 58.3 points, $42 / month (MiniRouter token estimate)
  • GPT-6 Astra: 54.8 points, $210 / month (MiniRouter token estimate)
  • Gemini 4 Hydrogen: 50.8 points, $32 / month (MiniRouter token estimate)
  • Claude Opus 5: 48.1 points, $105 / month (MiniRouter token estimate)
  • Claude Fable 5.1: 47.2 points, $210 / month (MiniRouter token estimate)
  • Grok 4.6: 44.3 points, $34 / month (MiniRouter token estimate)
  • Kimi K3: 43.6 points, $63 / month (MiniRouter token estimate)

Up to eight top-scoring matches with a cost estimate. The table has every match.

Your selection

Claude Fable 5.2

anthropic/claude-fable-5.2
Try in ChatModel details

61.8 Intelligence Index points. Scores describe the reported evaluation, not your prompts.

Context
1,000,000 tokens
Tools
Listed in catalog
Open weights
Unknown
Evaluation & evidence

Reported evaluation: Claude Fable 5.2 (High, Default Fallback) ↗

Artificial Analysis Intelligence Index · Version 5

Retrieved: . Measured: Unknown.

Score source: Source record ↗

Compatible API settings: Unknown. The reported evaluation settings cannot be mapped to a supported catalog configuration.

API key setup: create a key at Get API key, then set MINIROUTER_KEY to that API bearer key in your shell. Keep it private. Chat and the copied example use the selected model; they do not reproduce the benchmark settings. Trying the model bills normal usage.

Selected / month
$168.00

Same token volume at current MiniRouter rates. Reasoning tokens, cache writes, request minimums and routing can change the bill.

How this comparison works

Scores come from Artificial Analysis. Each task uses its own published index. Missing results stay unscored; benchmark coverage is not a measure of availability.

Only current MiniRouter catalog models appear.

Cost estimates use your input, output, and cache mix with MiniRouter’s catalog rates. Cache discounts apply only where a rate is published. Scores and costs come from different sources; an index point difference does not predict a percentage change in real-world quality.

Highest task benchmark means the highest reported selected-task score among matching models. Lowest estimated cost considers only matching models with a score and an available estimate. Ties retain the existing score, cost and name ordering. Neither is a recommendation based on measured quality per dollar.

Each evaluation’s evidence lists its task index, retained version and retrieval date. Unknown or different versions may not be directly comparable. Retrieval does not establish when a score was measured. Reported evaluation names describe third-party settings; catalog-compatible settings are not verified benchmark requests or provider reproductions.

First-party application quality and cost per acceptable result are unavailable. The sample workloads are authored assumptions, not observed customer usage. Try representative prompts before committing to a model.