Jais 3 vs MiniMax API

Jais 3

Jais 3 runs on the owner's hardware through OpenAI-compatible servers in vLLM and SGLang, with tool calls, three reasoning levels and a 512K context, from a 2.2 GB 0.9B model to a 758 GB 375B one. The weights download without sign-up under Apache-2.0, but every launch command trusts remote code, and no technical report or security contact is published.

MiniMax API

M3.1 Flash costs $0.15 in and $0.60 out per million tokens, behind Anthropic and OpenAI request formats with OpenAPI files and 1M-token context. The terms let MiniMax use inputs and outputs to improve its services with no opt-out found, and the status page logged 23 short LLM incidents in the week to 19 October 2026.

Jais 3

1. Start vLLM or SGLang with --reasoning-parser k2_horizon, and add --tool-call-parser k2_horizon (vLLM also --enable-auto-tool-choice) for tools, or tool calls come back as raw text
2. Pass reasoning_effort (high, medium or low) and tool_call_format inside chat_template_kwargs. Other values return a template error
3. Send each earlier assistant turn back with reasoning_content as a string, even an empty one. A missing or null thinking field makes the template raise an error
4. Send your own system message, or set default_system_prompt to an empty string. Without one the template inserts an identity prompt naming IFM, Inception42 and Cerebras
5. Budget at least 32,768 output tokens at high effort, which the cards recommend for every request

MiniMax API

1. Use the Anthropic-format base URL https://api.minimax.io/anthropic for tool calls. The OpenAI-format schema has no tool_choice or response_format
2. Set output_config.effort (or reasoning_effort) below max for routine calls. M3.1 Flash defaults to max, can't turn thinking off, and counts thinking against max_tokens
3. Don't rely on stop_sequences or top_k. The Anthropic-format endpoint ignores them
4. Back off on 429 and 529 yourself. No Retry-After header is documented
5. Use a pay-as-you-go key for production. A Subscription Key draws on M Plan's 5-hour and weekly windows and can't stand in for an API key