Contact, if you are interested in this website / domain name / sponsorship / advertisement / partnership
LLModel
Get expert help

LLM rankings you can tune

Our value index scores current models on capability tier, price, context, openness, input types and recency. The weights are yours: move the sliders or pick a preset, and the ranking updates instantly.

Priced, current models onlyData checked October 10, 2026Methodology

What matters to you?

40
25
10
10
5
10

Ranking

  1. MiniMax-M3MiniMax, Advanced, $0.30 / $1.2080
  2. DeepSeek-V4.1-FlashDeepSeek, Advanced, $0.30 / $1.2080
  3. GLM-5.3Z.ai (Zhipu AI), Frontier, $1.40 / $4.4079
  4. DeepSeek-V4-ProDeepSeek, Frontier, $1.32 / $3.9678
  5. Kimi K3Moonshot AI, Frontier, $3 / $1578
  6. GLM-5.3-FlashZ.ai (Zhipu AI), Efficient, $0.15 / $0.5075
  7. NVIDIA Nemotron 3 UltraNVIDIA, Advanced, $0.50 / $2.2074
  8. Mistral Large 4Mistral AI, Frontier, $1.36 / $4.1872
  9. Qwen3.8-2.4T-A95BAlibaba (Qwen), Frontier, $2 / $671
  10. Qwen3.8-MaxAlibaba (Qwen), Frontier, $2 / $669
  11. Gemini 3.8 FlashGoogle, Advanced, $0.75 / $3.7568
  12. Gemini 3.1 ProGoogle, Frontier, $2 / $1268
  13. NVIDIA Nemotron 3.5 Lightning 30B-A3BNVIDIA, Compact, $0.05 / $0.2068
  14. Gemini 3.6 FlashGoogle, Advanced, $0.75 / $3.7568
  15. Claude Opus 5.5Anthropic, Frontier, $4 / $2067
  16. Muse Spark 1.3Meta, Advanced, $1.25 / $4.2567
  17. Grok 4.7xAI, Frontier, $2 / $667
  18. Claude Haiku 5.5Anthropic, Efficient, $0.10 / $0.5066
  19. Gemma 4 31BGoogle, Efficient, $0.20 / $0.4066
  20. gpt-oss-120bOpenAI, Advanced, $0.15 / $0.6065
  21. Mistral Large 3Mistral AI, Advanced, $0.50 / $1.5065
  22. GPT-6 LunaOpenAI, Efficient, $0.10 / $0.5065
  23. Mistral Small 4Mistral AI, Efficient, $0.15 / $0.6064
  24. Mistral Medium 3.5Mistral AI, Advanced, $1.50 / $7.5063
  25. Claude Fable 5.1Anthropic, Frontier, $10 / $5063

Methodology

The index scores each current model with a published list price from 0 to 1 on six factors, multiplies each score by your weight, and rescales the total to 100. Legacy models and models without a confirmed price are left out so that every model is scored on the same inputs.

  • Capability tier. Frontier scores 1.0, advanced 0.75, efficient 0.5 and compact 0.3, plus 0.05 for a reasoning mode. Tiers are our editorial grouping of provider positioning and public benchmark standings; they are not our own test results.
  • Low price. Blended price (three parts input to one part output) on a log scale: the cheapest model scores 1, the most expensive 0. A log scale stops a single very expensive model from flattening everyone else.
  • Context window. Maximum context on a log scale between the smallest and largest window in the set.
  • Open weights. 1 if you can download the weights, 0 if not. License terms vary, so read them on the model page.
  • Inputs. One third of a point for each non-text input type (images, audio, video or PDF), capped at 1.
  • Newer release. Release date scaled between the oldest and newest model in the set.

The index is a shortlisting aid. It can't tell you which model writes the best answer for your task; only a test on your own data can. For raw capability, check the independent leaderboards below, which run their own evaluations.

Independent leaderboards and benchmarks

Read LLM benchmarks explained before you trust any single score.

Still choosing a model?

Send us your use case and constraints. We'll come back with a shortlist and an evaluation plan you can run on your own data.

  • Free and vendor-neutral: no model provider pays for placement.
  • Your details go only to partners you agree to.

Takes about 2 minutes. No sales calls unless you ask for one.

Frequently asked questions

What is the best LLM right now?

It depends on what you weigh. With the default balanced weights, the top of this index is a mix of frontier models and strong mid-priced ones. If price matters most, efficient models jump ahead; if you need to self-host, open-weight models lead. Independent benchmark leaderboards linked on this page track raw capability.

Do you run your own benchmarks?

No. The capability input is an editorial tier based on how providers position each model and where it stands on public benchmarks and arena rankings. Everything else in the index is calculated from published facts: list prices, context windows, licenses, input types and release dates.

Can a provider pay to rank higher?

No. Rankings come only from the formula on this page and the data on each model page. Sponsorships on LLModel are labeled and never change a score, an order or a verdict.