Mistral Saba
by Mistral AIChat
Mistral Saba is a 24B-parameter language model specifically designed for the Middle East and South Asia, delivering accurate and contextually relevant responses while maintaining efficient performance. Trained on curated regional datasets, it supports multiple Indian-origin languages—including Tamil and Malayalam—alongside Arabic. This makes it a versatile option for a range of regional and multilingual applications. Read more at the blog post here
- Context
- 32.8K tokens
- Max output
- 26,214 tokens
- Parameters
- 24B (total or active not stated)
- Released
- 2025-02-17
- Weights
- Closed
mistral/mistral-saba · also seen as mistralai/mistral-saba, mistral-saba-latest, mistralai/mistral-saba-2502
Specifications
Full facts as recorded, incl. what is unknown.
- Released
- 2025-02-17
- Weights
- Closed
- Input
- Text, PDF
- Output
- Text
- Context window
- 32,768 max across endpoints
- Max output
- 26,214
- Providers
- 2 + 3 gateways, not counted as providers
- Epoch Capabilities Index
- not evaluated
- Direct price range
- $0.200–$0.275 in · $0.600–$0.825 out per 1M tokens · across all 2 providers
More details
- Last updated
- 2025-02-17
- Knowledge cutoff
- 2024-09-30
- Parameters
- 24B (total or active not stated)
- Developer country
- France
Capabilities
Across all endpoints: supported if any endpoint supports it.
- Tool calling
- Parallel toolsunknown
- Structured output
- Reasoning
- Image input
- PDF input
- Audio input
- Audio outputunknown
- Video input
- Prompt caching
- Web searchunknown
- Computer useunknown
- System messagesunknown
- Assistant prefillunknown
- Temperature
- Streamingunknown
- Attachments
Endpoints
One row per provider and variant; gateways are listed separately.
Loading endpoints…
Price history
Atom feedNo change since 19 Oct 2026. The cheapest direct price has been $0.20 in / $0.60 out per 1M tokens at Mistral on every data refresh since then.
Tracked from 19 Oct 2026. Headline-eligible direct prices only, so a price the sanity check flags never draws a spike. Series as JSON
Cost calculator
Loading endpoints…
Similarly priced models
Blended rate per 1M tokens, each at its own cheapest provider (10M input : 2M output mix).
| Model | Blended/1M | In/1M | Out/1M | Context | vs Mistral Saba |
|---|---|---|---|---|---|
| Nemotron Nano 12B v2 VLNVIDIA | $0.267 | $0.20 | $0.60 | 131K | +1% |
| Agnes 1.5 ProSapiens AI (Agnes AI) | $0.267 | $0.16 | $0.80 | 256K | +1% |
| MiMo-V2.5-ProXiaomi | $0.267 | $0.20 | $0.60 | 1.05M | +1% |
| Doubao Seed 2.0 LiteByteDance Seed | $0.267 | $0.146 | $0.874 | 256K | +1% |
| Qwen3-Next 80B-A3B InstructAlibaba (Qwen) | $0.258 | $0.09 | $1.10 | 131K | −2% cheaper |
| K2-ThinkLLM360 | $0.255 | $0.17 | $0.68 | 128K | −4% cheaper |
| Phi-4-MiniMicrosoft | $0.255 | $0.17 | $0.68 | 131K | −4% cheaper |
| Qwen3.8 Flash NextAlibaba (Qwen) | $0.25 | $0.20 | $0.50 | 262K | −6% cheaper |
More from Mistral AI
Newest Mistral AI models in the catalogue. All Mistral AI models
| Model | Released | Context | Listed price |
|---|---|---|---|
| Mistral Large 4 | 2026-10-06 | 1.05M | from $0.68 |
| Voxtral Small 24B 2507 STT | 2026-08-13 | — | unpriced |
| Leanstral 1.5 | 2026-06-30 | 262K | unpriced |
| Mistral Code Latest | 2026-06-02 | 128K | from $0.30 |
| Mistral Code Agent Latest | 2026-06-02 | 256K | from $0.40 |
| Voxtral Mini Transcribe | 2026-05-15 | 16.4K | unpriced |
| Mistral Medium 3.5 | 2026-04-30 | 262K | from $1.50 |
| Mistral Medium 3.5 | 2026-04-29 | 262K | from $1.50 |