llmgauge

comparisons

Compare LLMs head-to-head

156 side-by-side comparisons across 24 leading models — token pricing, context window, modality and the real monthly cost for a typical workload. Pick a matchup, or start from a model.

Models

GPT-5.5 $30.00/1M · 1.1M GPT-5.4 $15.00/1M · 1.1M GPT-5.4 mini $4.50/1M · 400K GPT-5.4 nano $1.25/1M · 400K gpt-oss-120b $0.18/1M · 131K Claude Fable 5 $50.00/1M · 1M Claude Opus 4.8 $25.00/1M · 1M Claude Sonnet 4.6 $15.00/1M · 1M Claude Haiku 4.5 $5.00/1M · 200K Gemini 3.1 Pro $12.00/1M · 1.0M Gemini 2.5 Pro $10.00/1M · 1.0M Gemini 2.5 Flash $2.50/1M · 1.0M Gemini 2.5 Flash-Lite $0.40/1M · 1.0M Grok 4.3 $2.50/1M · 1M Grok 4.20 $2.50/1M · 2M DeepSeek V4 Pro $0.87/1M · 1.0M DeepSeek V4 Flash $0.28/1M · 1.0M Mistral Medium 3.5 $7.50/1M · 262K Mistral Large 3 $1.50/1M · 262K Llama 4 Maverick $0.60/1M · 1.0M Qwen3 Max $3.90/1M · 262K GLM 4.6 $1.74/1M · 203K Kimi K2 Thinking $2.50/1M · 262K MiniMax M2.5 $0.90/1M · 205K

Flagship matchups

Frontier models — quality-first, output ≥ $15/1M.

Mid-range matchups

The workhorse tier — output $2–15/1M.

Budget matchups

High-volume, cost-sensitive — output under $2/1M.

How these comparisons work

Every comparison is generated from the same sourced catalog the rest of the site uses — no hand-picked numbers. Each page shows input and output token prices, the context window, accepted modalities, the price source, and the modelled monthly cost at a typical workload, with the cheaper option marked. Prices are list or routed market figures captured with a date; the ranking can shift with your own token mix, which is why every page links to the cost calculator.

Related