Free Intelligence · Model access

Mistral & Grok (via Grok Bot)

Two non-Chinese model families that sit next to the OpenCode Go & Command Go tracker: what each offers, what's free, what it costs, and when to pick it. Every figure links to the vendor's own page and was checked Oct 16, 2026.

Mistral

Free

  • Free plan (Vibe app, web & mobile): limited messages, web searches and coding sessions; includes Vibe for code in the terminal and Mistral Studio.
  • API free mode: API keys on by default, no credit card, included monthly usage with the lowest rate limits. Meant for evaluation and prototyping.

mistral.ai/pricing · docs: activate Studio · docs: usage & limits · checked Oct 16, 2026

Paid plans

  • Pro $14.99/mo: all-day coding in CLI, IDE or web, higher limits, $15/mo in API credits.
  • Students: Pro for $5.99/mo.
  • Team $24.99 and Enterprise (custom).
  • API pay-as-you-go: prepaid credits; rate tiers rise with cumulative billing.

mistral.ai/pricing · help: rate limits · checked Oct 16, 2026

Current models & API prices (USD per 1M tokens, standard)

ModelUseLicenseInputCachedOutput
Mistral Medium 3.5Mistral's pick for most tasks and codingModified MIT$1.50$0.15$7.50
Mistral Large 4Open-weight general multimodalOpen weight$0.68 sale (list $1.36)$0.07$2.09 sale (list $4.18)
Mistral Large 3Previous flagshipApache 2.0$0.50$0.05$1.50
Mistral Small 4Cheap hybrid: instruct, reasoning, codingApache 2.0$0.15$0.015$0.60
Ministral 3 14B / 8B / 3BSmall text + visionApache 2.0$0.20 / $0.15 / $0.10—$0.20 / $0.15 / $0.10
Codestral (25.08)Code completion (FIM)—$0.30$0.03$0.90

Prices: docs.mistral.ai/inference/pricing · models & licenses: models overview · checked Oct 16, 2026. Batch is 50% off; cached input up to 90% off. Large 4 "sale" prices are as shown on Mistral's page that day.

Coding models: Codestral is current for completion. Devstral is retired: Devstral 2 was deprecated May 22, 2026 and retired July 31, 2026, with Mistral Medium 3.5 named as the replacement. Mistral's agentic coding is now Vibe for code (Free and Pro plans). Source, checked Oct 16, 2026.

Grok, and Grok via Grok Bot

What Grok Bot is

Grok Bot is xAI's agent app: you hand a bot a task and it works through connected tools. Connect X in Grok Bot and it creates an X developer account if you don't have one. Paid Grok Bot users get free X API credits to start. Bots can search posts, read timelines, check mentions and summarize what's happening on X.

x.ai/news/grok-bot-and-x · checked Oct 16, 2026

xAI API access

  • Sign up at console.x.ai, load prepaid credits, create a key; base URL https://api.x.ai/v1 (OpenAI/Anthropic-compatible).
  • Free: the Console Playground comes with every account, for testing models before you pay.
  • Free API credits: none currently listed. The $25/mo free credits were a public-beta offer that ended with 2024.

x.ai/api · quickstart · beta announcement · checked Oct 16, 2026

Grok text models (USD per 1M tokens, prompts < 200K tokens)

ModelContextInputCachedOutput
grok-4.7 (xAI's pick for code and chat)500K$2.00$0.50$6.00
grok-4.6500K$2.00$0.50$6.00
grok-4.5500K$2.00$0.30$6.00
grok-4.31M$1.25$0.20$2.50
grok-4.20-0309 (reasoning / non-reasoning / multi-agent)1M$1.25$0.20$2.50
grok-build-0.1256K$1.00$0.20$2.00

docs.x.ai/docs/models · checked Oct 16, 2026. Prompts of 200K tokens or more bill at 2× for the whole request. Images from $0.02 each, video from $0.02/sec, voice $0.08/min.

Which to use when

If you want…UseWhy
The cheapest paid coding planCommand Go$1/mo for $10 of credits ($3/5h, $6/week). CLI only, no API. Also includes Mistral Large 4 and Grok 4.5.
High-volume open models in any agentOpenCode Go$10/mo (Go Plus $40) with large per-model allowances; works with OpenCode or any agent.
Free chat + terminal coding without a cardMistral FreeVibe app plus Vibe for code, and an API free mode for prototyping.
Cheap API calls, EU vendor, open weightsMistral Small 4 / Large 4$0.15/$0.60 (Small 4); Apache-2.0 and open-weight options you can self-host.
Code completion in an IDECodestralPurpose-built for completion; $0.30/$0.90.
Frontier reasoning or 1M contextGrok 4.7 / Grok 4.3 (xAI API)Grok 4.7 is xAI's top model; 4.3 and 4.20 give 1M context at $1.25/$2.50.
An agent that does the work on X and other toolsGrok BotConnectors (including X with starter API credits for paid users) instead of raw API calls.

Plan figures from commandcode.ai and opencode.ai/go, the other figures from the sources above. All checked Oct 16, 2026.

Route notes (how we use them)

From our internal model-routing checklist. No keys or account details are published.