MiniMax Code Plans & Cerebras Code Plans: The BEST ALTERNATIVES to CLAUDE CODE just ARRIVED!

Key Concepts

• Minimax M2: A new, highly-rated open-weight AI model, currently ranked second on the speaker's Aentic benchmarks.
• API Pricing: Minimax's API will have new pricing starting tomorrow, with input tokens costing $0.3 per million and output tokens costing $120 per million.
• Coding Plans: Minimax is launching tiered coding plans ($10, $20, $50 per month) offering usage limits and cost savings compared to other providers like Anthropic.
• Anthropic Claude Code Max: A benchmark for comparison, with Minimax's plans offering significantly lower costs and higher usage limits for equivalent tiers.
• GLM 4.6: Another model mentioned, with reliability issues and slower performance compared to Minimax M2.
• See Code Plan: A $50 plan offering access to GLM 4.6 with high token limits per minute and day, and a $200 plan with even higher limits and speed.
• Cursor Composer: A tool mentioned for comparison with See Code plan's performance.
• OpenAI Compatible API: Minimax's API is designed to function similarly to OpenAI and Anthropic APIs.
• Token Speed: Minimax API is currently delivering approximately 97 tokens per second for free.

Minimax New Coding Plan and API Pricing Analysis

This video discusses the recent launch of Minimax's new coding plan and API pricing structure, evaluating its value proposition against existing offerings. The speaker highlights the Minimax M2 model as a strong performer, currently ranking second on their Aentic benchmarks and preferred over GLM 4.6.

Minimax M2 Model and API Pricing

• Model Performance: Minimax M2 is described as an "awesome model" with open weights, currently outperforming GLM 4.6 for the speaker. The speaker notes that two new open-weight models, including a smaller model and a vision model, are expected soon, which could alter the landscape.
• API Cost: The current API pricing for Minimax M2 is $0.3 per million input tokens and $120 per million output tokens. This is significantly cheaper than other models, with the speaker estimating it to be about 13 times cheaper than Anthropic's Sonnet, while delivering comparable or better results.
• Cost-Effectiveness: The speaker emphasizes the affordability, suggesting that even without a plan, usage might not exceed $20-$30. This makes it suitable for long sessions without the high cost associated with models like Claude.
• Reliability and Speed: The Minimax API is noted for its reliability and consistent speed, currently delivering around 97 tokens per second even while free and under heavy usage. This is contrasted with GLM 4.6, which has experienced reliability issues and slower performance.

Minimax Coding Plans

Minimax is introducing three coding plans with varying monthly costs and usage limits:

• $10 Per Month Plan:
  • Comparison: Equivalent to Anthropic's Claude Code Max 5x plan.
  • Cost: 10% of Claude's plan cost.
  • Usage Limit: 2 times the usage limit of Claude's plan, allowing at least 500 messages every 5 hours.
• $20 Per Month Plan:
  • Comparison: Equivalent to Anthropic's 20x plan.
  • Cost: 10% of Claude's plan cost.
  • Usage Limit: Approximately 1.5 times the limit of Claude's plan, providing at least 1,300 messages every 5 hours.
• $50 Per Month Plan:
  • Comparison: Offers 5 times the usage of Anthropic's $200 Claude Code Max 20x plan.
  • Usage Limit: At least 4,500 messages every 5 hours.

The speaker finds these plans "insane" in terms of value and pricing, although they would have preferred a lower entry-level plan like GLM's $3 option. The pricing is seen as a positive development driven by models from China, pushing the frontier with cost-effectiveness in mind. These plans are expected to be compatible with various tools and function like OpenAI or Anthropic compatible APIs.

See Code Plan and GLM 4.6

• See Code Plan ($50/month): Provides access to GLM 4.6 with significant token limits: 1 million tokens per minute and 24 million tokens daily.
• See Code Plan ($200/month): Offers even higher limits, including approximately 1,000 tokens per second for GLM, described as "insane" and significantly faster and better than Cursor Composer.
• API Pricing for See Code: The API for this plan costs $2.50 per million input tokens and $2.75 per million output tokens, which is also considered "great" and offers faster speeds than Cursor Composer at a lower cost.
• Availability Issue: The speaker notes that the See Code plan is currently out of stock.
• Future Integration: The speaker hopes for Minimax M2 to be integrated into the See Code plan for potentially even faster performance due to its smaller size.

Future AI Landscape

The speaker anticipates a significant year for AI, with the upcoming releases of GPT 5.1 and Gemini this month, following a less exciting previous year.

Call to Action

The speaker invites viewers to share their thoughts, request performance tests and tutorials for the new coding plan API, and subscribe to the channel. Options for supporting the channel through Superthanks or channel memberships are also mentioned.