DeepSeek flagship 1.6T MoE with 49B active, 1M context, MIT weights. API $0.29/$1.20 cached $0.03.

OpenAI GPT-5.6 Luna: fast affordable GPT-5.6 tier, 128k context, $1/$6 per 1M tokens.

OpenAI GPT-5.6 Terra: balanced daily driver in the 5.6 family, 400k context, $2.50/$15 per 1M tokens.

OpenAI frontier flagship GPT-5.6 Sol: 1.05M context, effort up to max plus ultra, $5/$30 per 1M tokens.

OpenAI GPT-6 Luna: the cheapest GPT-6 tier at $0.10/$0.50 per 1M tokens with a 1.05M context window and max-effort reasoning.

OpenAI GPT-6 Sol: 1.05M context, reasoning up to max effort and $2/$10 per 1M tokens, with day-one GitHub Copilot support.

SpaceXAI's frontier 1.5T reasoning model with 500K-token context, agentic coding strength, and real-time X data at roughly half the price of rivals.

SpaceXAI's Grok 4.7 flagship for coding and agents: 500K context, $1.60/$4.80 per 1M tokens, 46.4 on the Artificial Analysis index.

xAI / SpaceXAI Grok family. Current docs flagship is grok-4.6: 500k context, $2/$6 per 1M tokens, plus Imagine image/video and Voice APIs.

OpenAI's previous frontier model: 400k context, 128k max out, $1.75/$14 per 1M tokens. Docs now recommend GPT-5.6 Sol.

Anthropic's Mythos-class flagship made safe for general use, with 1M-token context, frontier coding and vision, days-long agentic autonomy, and $10/$50 per 1.

Anthropic's previous Opus snapshot: 200k context, $5/$25 per 1M tokens. Later billed lines include Opus 4.6 and Opus 5.

Anthropic's frontier Opus model with 1M-token context and effort-controlled reasoning, near-Fable-5 intelligence at Opus-4.8 pricing for agents and coding.

Previous DeepSeek-R1 snapshot. Not an o1 / 2026 reasoning default.

Previous Gemini 2.0 Flash Thinking snapshot. Not a 2026 o1-mini alternative.

NVIDIA's efficient open-weight 30B MoE hybrid model with 3B active parameters, 1M-token context, and single-GPU deployment for local reasoning and coding.

Previous xAI Grok Beta snapshot. Not a 2026 reasoning default.

Previous o1-mini snapshot. Not the 2026 reasoning default.

Previous o1-preview snapshot. Not a cheaper o1 and not 2026 default.

Previous QwQ-32B-Preview snapshot.

Evaluate research rigor systematically - assess methodology, experimental design, statistical validity, biases, confounding, and evidence quality using GRADE.

Jeff is an open-source family of 0.8B and 2B decision models that return a calibrated probability per option in about 22 ms on a single GPU.