One agent session
1M input tokens and 100K output tokens, no caching.
- Input1M x $0.20 = $0.20
- Output0.1M x $0.40 = $0.04
$0.24per session
AI21 Labs - Open weights
Jamba2 Mini is an open-weights model from AI21 Labs that costs $0.20 per 1M input tokens and $0.40 per 1M output tokens on the first-party API, with a 256K-token context window.
Last verified from ai21.com. All sources
| Item | Price / 1M tokens |
|---|---|
| Input | $0.20 |
| Output | $0.40 |
Note: The pricing page lists this model as "Jamba Mini" with no version; the models page says the jamba-mini alias currently points to jamba-mini-2-2026-01 (Jamba Mini version 2, 52B total / 12B active parameters), and jamba-mini-2 points to the same snapshot. The older Jamba Mini 1.7 (jamba-mini-1.7-2025-08) was deprecated on 2026-02-01. context window is stated as "256K" on the models page and in the launch post, recorded here as 256000. max output comes from the chat API reference: max_tokens allows at most 4096 for Jamba models. Text in, text out. AI21 publishes no cache or batch prices. release date is the publish date of the Introducing Jamba2 launch post (article:published_time 2026-01-08). it has no Vercel AI Gateway id: ai-gateway.vercel.sh returned model_not_found for ai21/jamba-mini, and no Jamba entry turned up on the gateway pages read.
jamba-mini-2Worked from the standard list rates above, with the math shown so you can swap in your own token counts.
1M input tokens and 100K output tokens, no caching.
$0.24per session
10,000 requests a day at 2K input and 500 output tokens each.
$6.00per day
$180.00 per 30-day month
| Model | Input / 1M | Output / 1M |
|---|---|---|
| Jamba2 Mini (this page) | $0.20 | $0.40 |
| Gemini 2.5 Flash-LiteGoogle | $0.10 | $0.40 |
| GLM-4.7-FlashXZ.ai | $0.07 | $0.40 |
| GPT-4.1 nanoOpenAI | $0.10 | $0.40 |
| GPT-5 nanoOpenAI | $0.05 | $0.40 |
| Jamba Large 1.7AI21 Labs | $2 | $8 |
Jamba2 Mini costs $0.20 per 1M input tokens and $0.40 per 1M output tokens on AI21 Labs's first-party API, per ai21.com on October 17, 2026.
Jamba2 Mini has a 256,000-token context window and returns up to 4,096 output tokens per response.
Yes. Jamba2 Mini's weights are published at https://huggingface.co/ai21labs/AI21-Jamba2-Mini under the Apache 2.0.
AI21 Labs released Jamba2 Mini on January 8, 2026.
Last verified from ai21.com.
Newsletter
One email per week. New model releases, price cuts and what they mean for your bill.