Refreshed October 4, 2026 · 19 models

Every frontier model. One place.

Compare 19 production models across OpenAI, Anthropic, Google, SpaceXAI (xAI), Moonshot, Alibaba, Z.ai, Xiaomi, DeepSeek and MiniMax — on benchmarks, pricing, context, and practical task fit. Pick a model in 30 seconds, or evaluate a real task in two minutes.

Pick a model in 30 seconds Browse all models

Sep · New line GPT-6 Astra / 6.1 Sol / 6 Luna OpenAI's GPT-6 family. Astra is the $10/$50 flagship (Sep 3); GPT-6.1 Sol nearly matches it at $2/$10 with 95%-off cache reads (Sep 29); Luna resets the budget lane at $0.10/$0.50.

Sep · New GA Claude Opus 5.5 + Sonnet 5.5 + Fable 5.1 Opus 5.5 is 20% cheaper than Opus 5 ($4/$20) and #1 on the Artificial Analysis index (58). Sonnet 5.5 holds $2/$10. Fable 5.1 cuts cache reads 75%.

Aug–Sep · Open weights MiMo-V2.6-Pro, GLM-5.3, Qwen3.8 The open-weight lead changed hands: MiMo-V2.6-Pro (46), GLM-5.3 and Qwen3.8-Max (45) now beat Kimi K3 (44). DeepSeek V4.1 Flash is $0.30/$1.20, with 2× peak-hour pricing.

Quick pick

Three questions — one recommended model. For nuanced tradeoffs, use the full evaluator below.

1. What's the work?

2. Budget?

3. Speed need?

The full lineup

Click any card to see strengths, watch-outs, and tooling. Pin up to 3 to compare side-by-side.

Task evaluator

Paste a real task, choose the dominant work type, and tune cost, speed, and context sensitivity. Scored against all 19 production models.

Your task

Primary task type

Approximate input size

Cost sensitivity

Latency sensitivity

Context importance

Tool-use importance

Recommendation

Relative fit score across all 19 production models — higher is better, and keyword matches in your task can push it above 100. Top 5 shown with tradeoff rationale.

Side-by-side prompt runners

Top 3 recommended models get starter prompts. Paste into vendor consoles or OpenRouter and score outputs consistently.

Scorecards

Repeatable rubric for head-to-head runs.

Best team of models

Splits a workflow into subtasks and assigns the best model for each slice — routing across the full 19-model set.

Multi-step workflow

Routed plan

Opinionated defaults — override freely.

Ones to watch

Models that are announced, restricted or still in preview — plus two open-source agent harnesses you can adopt now. Worth tracking closely.

How to use this in practice

Start with the quick pick. For zero-cost experiments, try Qwen3.8-27B (free OpenRouter tier) or self-host Gemma 4 / GLM-5.3-Flash before paying for frontier models, and use the router for compound workflows. Compare cost per finished task, not per-token price — Grok 4.7 and Sonnet 5.5 at max effort show how the same sticker price can double the bill. Watch: Gemini 4 Argon and Claude Haiku 5.5 are the next likely shake-ups; near-frontier pricing has converged at ~$2/$10 (GPT-6.1 Sol, Sonnet 5.5), and the open-weight lead now sits with MiMo-V2.6-Pro, GLM-5.3 and Qwen3.8.