Skip to content

Craft & SystemsModels & Failure

AI Model Reference Guide

Compare current AI model families, understand reasoning and sampling controls, and choose models for research reasoning, writing, and analysis tasks.

The available models change fast; the task-fit question stays constant.

Overview: current, but not permanent

Model names, aliases, pricing, rate limits, and control parameters shift frequently, and a provider's latest release does not immediately appear in every application menu. Before building a workflow around a specific model ID, check the provider's official documentation and your client's supported route.

The practical question is not which model tops a benchmark. It is which model, access path, and control surface fit the work in front of you:

  • reasoning depth: can the model handle theoretical tension and multi-step analysis?
  • context capacity: can it read the relevant corpus without destructive chunking?
  • writing behavior: does it preserve academic nuance and voice?
  • cost and latency: can you afford to iterate across drafts or batches?
  • control surface: does it respond to reasoning effort, thinking level, prompt constraints, or sampling parameters?

Current model families

OpenAI GPT Family:

  • Current provider-native choices include gpt-6-astra, gpt-6.1-sol, gpt-6.1-terra, and gpt-6-luna. The GPT-5.6 family remains available as gpt-5.6-sol, gpt-5.6-terra, and gpt-5.6-luna; the gpt-5.6 alias routes to Sol, not Astra.
  • Context: roughly 1M on Astra, Sol, Terra, and Luna. Check the model catalogue (opens in a new tab) for exact limits.
  • Control surface: reasoning.effort where supported. Astra and GPT-6.1 Sol support low through max, but not none; Sol defaults to medium. GPT-6.1 Terra also supports none. Tool calling on these three models requires the Responses API. Check the exact model's controls before reusing an older configuration.
  • Suggested collaboration: work with Astra or GPT-6.1 Sol on complex research and coding, and with Luna on volume tasks. Keep older Terra where it already fits, or compare GPT-6.1 Terra (opens in a new tab) for visual reasoning and writing. Compare Sol with Astra on the same task, as OpenAI's model selection guide (opens in a new tab) recommends.
  • Source: OpenAI latest-model guide (opens in a new tab).

Anthropic Claude Family:

  • Current provider-native IDs include claude-fable-5-2, claude-opus-5-5, claude-sonnet-5-5, and claude-haiku-5-5. Fable 5.1 (claude-fable-5-1) and Haiku 4.5 remain active. Haiku 5.5 is the current faster, lower-cost option.
  • Context: 1M on Fable 5.2, Fable 5.1, Opus 5.5, and Sonnet 5.5. Fable 5.2 allows up to 128K output.
  • Control surface: Fable 5.2 always uses adaptive thinking, with default high effort. Omit temperature, top_p, and top_k; forced tool choice (any or a named tool) is unsupported. Check each Claude model's controls separately.
  • Fable 5.2 data handling: the API requires 30-day retention unless Anthropic expressly authorizes zero data retention. Check this condition before sending sensitive material; see the Fable 5.2 guide (opens in a new tab).
  • Suggested collaboration: start defined analysis, writing, or coding tasks with Opus 5.5 or Sonnet 5.5. Keep Fable 5.1 where it works, and compare Fable 5.2 for multi-agent coordination.
  • Haiku 5.5 offers a 1M context window, up to 128K output, and adaptive thinking on by default with medium effort. Its pricing (opens in a new tab) changes above 100K input tokens; see the Haiku model guide (opens in a new tab) before migrating manual-thinking settings from 4.5.
  • Lifecycle: check Anthropic's model deprecations guide (opens in a new tab) before relying on older 4.5 aliases stored in an application menu.
  • Source: Claude model docs (opens in a new tab).

Kimi Family (Moonshot AI):

  • Current flagship provider-native ID: kimi-k3.1-pro. Kimi K3 remains supported as kimi-k3.
  • Context: 1M on Kimi K3.1 Pro and K3; 256K on K2.7 Code and K2.6.
  • Access and lifecycle: new accounts cannot access the K2 direct API. Accounts registered before November 1 can continue, with deprecation as expected in late December 2026. Check the Kimi model list and lifecycle (opens in a new tab) before planning a new direct-API route.
  • Direct API access: K3 requires a successful top-up of at least $1. The same documented minimum applies to K3.1 Pro. Account tier affects rate limits, and API billing is separate from Kimi Code subscription access.
  • Control surface: Kimi K3 always thinks and accepts reasoning_effort values low, high, and max (default max). Omit sampling parameters.
  • Suggested collaboration: We work with Kimi as a regular collaborator across reasoning, writing, coding, and sustained project work. Work with Kimi K3 or K3.1 Pro as a lead analyst or coordinator when you want an independent framing from the start. See a workflow that starts with Kimi.
  • Sources: Kimi model list and lifecycle (opens in a new tab), Kimi K3 quickstart (opens in a new tab).

Google Gemini Family:

  • Current choices: Google's latest guide positions gemini-4-hydrogen as its general-purpose starting point, gemini-4-carbon for its most demanding reasoning, and gemini-4-argon as an efficient frontier-reasoning option. gemini-3.8-flash remains a stable Flash choice; gemini-3.6-flash remains listed; gemini-3.5-flash-lite serves volume work; gemini-3.1-pro-preview remains in preview. Google has deprecated gemini-3.7-flash and routes its requests to 3.8 (opens in a new tab).
  • Context and output: Argon (opens in a new tab), Hydrogen (opens in a new tab), and Carbon (opens in a new tab) share a 1M-token context window across input, thinking, and output. Synchronous output caps at 64K tokens; longer outputs use the long-decode route.
  • Control surface: omit temperature, top_p, and top_k on Gemini 4 and current Gemini 3.x models. Use thinking_level: Hydrogen defaults to high thinking and adds xhigh. Carbon and Argon offer low, medium, and high (defaulting to medium), and Argon does not support explicit context caching. Gemini 3.8 Flash does not support minimal.
  • Access and pricing: Argon is available on the paid Gemini API and AI Studio with no Free tier. Check current Gemini pricing (opens in a new tab) before running large output budgets.
  • Suggested collaboration: Gemini 3.8 Flash is a dependable starting point for large-corpus synthesis and high-volume screening; compare Gemini 4 Argon, Carbon, or Hydrogen where your client supports them.
  • Source: Gemini latest-model guide (opens in a new tab).

Access paths

Provider-native IDs, gateway-prefixed slugs, client aliases, and subscription menus are distinct surfaces. For example, GPT-5.5 retired from Codex with ChatGPT sign-in (opens in a new tab) on October 14, 2026, while its OpenAI API availability and API-key-authenticated Codex sessions were unaffected.

Access pathBest forNotes
Provider web appFast explorationGood for testing a model before wiring API access.
Google AI StudioGemini API keys and selected free-tier experimentsArgon is paid. Use for direct Gemini model access, not Antigravity CLI auth.
Cherry StudioGUI model comparison and knowledge basesVisual desktop workspace for comparing models side by side.
OpenCodeCLI-side provider comparisonHolds the terminal harness constant while comparing models across providers.
Vox MCPMulti-model access inside an MCP clientPure passthrough; good for challenge and triangulation.
Claude Code / Codex / Antigravity CLIWorkspace agent workProject-folder workflows with product sign-in or API keys.

Configuration: defaults first

Sampling is no longer the default lever

Tuning temperature on every model is an outdated habit. Most current reasoning models either reject sampling overrides, ignore them, or perform best at provider defaults.

Omit temperature, top_p, and top_k unless the provider documents that the model supports and benefits from changing them:

Provider / modelRecommendation
Claude 5 familyKeep sampling defaults; use effort and the model's documented thinking controls.
Gemini 4 and 3.xSet the model's documented thinking_level; omit deprecated sampling parameters.
Kimi K3 / K2.7 / K2.6Omit sampling parameters; use reasoning_effort on K3.
GLM-5.3Always-on thinking; use reasoning_effort (low, high, or max).
DeepSeek V4Use thinking mode and reasoning_effort; keep other controls at provider defaults.
xAI Grok 4.7Prefer documented reasoning controls; avoid incompatible parameters on reasoning models.

If a non-reasoning model requires a temperature and the provider offers no task-specific guidance, start at 1.0 and lower it only for deterministic extraction or classification.

Working across models

Pick each model for the role it plays in a differentiated team or single-agent workflow. We work with Kimi as a regular collaborator across reasoning, writing, coding, and sustained project work because its perspective differs from Claude and GPT from the first framing onward. A model's role can shift with the stage of the project.

Work shapeGood starting modelsWhy
One bounded theory, critique, or writing taskClaude Opus 5.5 or Claude Sonnet 5.5Strong nuance and close adherence to a scoped brief.
Carry out an agreed project planGPT-6 Astra or GPT-6.1 SolFollows through reliably across multi-step plans once criteria and boundaries are explicit.
Develop an independent framing from the outsetKimi K3 or Kimi K3.1 ProBrings a distinct perspective across reasoning, writing, coding, and sustained project work.
Multi-model coordinationClaude Fable 5.1, GPT-6 Astra, or Kimi K3; compare Fable 5.2 where its retention rules fitCoordinates specialist passes or alternates between delegation and direct work.
Large literature sets and long-context synthesisGemini 3.8 Flash; compare Gemini 4 Argon or GLM-5.3Handles large corpora cleanly before you narrow to close reading.
Lower-cost volume screeningDeepSeek Flash, Gemini 3.5 Flash-Lite, GPT-6 Luna, or Qwen3.7-Plus / Qwen3.8-FlashKeeps iteration affordable on screening, tagging, and extraction; compare error rates alongside cost.
Quick Claude-family checksClaude Haiku 5.5Fast, lower-cost option for bounded checks while Haiku 4.5 remains in legacy menus.
Privacy-sensitive or offline corporaLocal / open-weight modelsKeeps restricted material on your own hardware.

In our daily runs, Claude Opus 5.5 sometimes pauses mid-task to re-confirm actions even after permissions and boundaries are granted, whereas GPT-6 Astra and GPT-6.1 Sol tend to carry an agreed plan through.

Free and low-cost options

Google AI Studio offers free access to selected Gemini models, though Argon is paid-only. Check Gemini pricing (opens in a new tab) and rate limits (opens in a new tab) before relying on a quota for batch work.

Other ways to manage cost:

  • Use OpenCode or Cherry Studio to compare providers on a small sample before running a full corpus.
  • Work with DeepSeek or Qwen on high-volume screening when their accuracy on your sample matches your rubric.
  • Work with local models when data governance matters more than frontier capability.
  • Reserve higher-cost reasoning models for the synthesis and critique stages where their judgment pays off.

Next step

Test these candidates against your own materials with the AI Model Discovery Protocol.

cite this page

Lin, X. (2026). AI Model Reference Guide. Research Memex. https://research-memex.org/docs/craft/ai-model-reference-guide

@misc{docs-craft-ai-model-reference-guide-2026,
  author = {Xule Lin},
  title = {AI Model Reference Guide},
  year = {2026},
  howpublished = {\url{https://research-memex.org/docs/craft/ai-model-reference-guide}},
  note = {ORCID: 0000-0001-7885-4194}
}

one renderingthe source remains