AI Models, ZAR Pricing & Benchmarks
Access state-of-the-art LLMs from OpenAI, Anthropic, Google, Meta, DeepSeek and more. All models are billed per million tokens in South African Rands (ZAR) via your prepaid wallet with zero offshore forex surcharges.
Available Models
50 models| Model | Provider | Context | Input / 1M | Output / 1M | Status | Specs & Actions |
|---|---|---|---|---|---|---|
|
M
meta-llama/llama-3.3-70b-instruct
primary — high intelligence, 70B open weights instruction-tuned model.
|
Meta Llama | 128.0 KB tokens | R 5.1075 | R 13.2796 | Healthy |
Specs →
|
|
D
deepseek/deepseek-chat
Fast, low-cost utility with strong coding and reasoning capabilities.
|
DeepSeek | 160.0 KB tokens | R 6.5734 | R 26.2706 | Healthy |
Specs →
|
|
O
openai/gpt-4o-mini
Lightweight, cost-efficient model optimized for structured JSON mode and function calling.
|
OpenAI | 125.0 KB tokens | R 3.8307 | R 15.3226 | Healthy |
Specs →
|
|
A
anthropic/claude-3.5-sonnet
Frontier-class intelligence for complex reasoning, code, and analysis.
|
Anthropic | 976.6 KB tokens | R 76.6129 | R 383.0644 | Healthy |
Specs →
|
|
M
meta-llama/llama-3.1-70b-instruct
Highly capable 70B open instruction model.
|
Meta Llama | 128.0 KB tokens | R 18.3871 | R 18.3871 | Healthy |
Specs →
|
|
M
meta-llama/llama-3.1-8b-instruct
Ultra-fast, cost-effective 8B instruction model.
|
Meta Llama | 128.0 KB tokens | R 1.2769 | R 2.0430 | Healthy |
Specs →
|
|
M
meta-llama/llama-3.2-3b-instruct
Edge and lightweight processing model.
|
Meta Llama | 128.0 KB tokens | R 1.2999 | R 8.5551 | Healthy |
Specs →
|
|
M
meta-llama/llama-3.2-1b-instruct
Compact high-throughput model.
|
Meta Llama | 58.6 KB tokens | R 0.6895 | R 5.1331 | Healthy |
Specs →
|
|
M
meta-llama/llama-4-scout
Next-generation Llama preview architecture.
|
Meta Llama | 320.0 KB tokens | R 2.5538 | R 7.6613 | Healthy |
Specs →
|
|
D
deepseek/deepseek-v4-flash
DeepSeek next-gen fast flash model.
|
DeepSeek | 1.0 MB tokens | R 2.3239 | R 4.6478 | Healthy |
Specs →
|
|
D
deepseek/deepseek-v4-pro
High-capacity DeepSeek V4 reasoning engine.
|
DeepSeek | 1.0 MB tokens | R 40.8602 | R 81.7204 | Healthy |
Specs →
|
|
D
deepseek/deepseek-v4.1-flash
Enhanced DeepSeek V4.1 with speed optimizations.
|
DeepSeek | 1.0 MB tokens | R 1.1492 | R 12.7688 | Healthy |
Specs →
|
|
D
deepseek/deepseek-v4-flash-vision-exp
DeepSeek multimodal visual reasoning preview.
|
DeepSeek | 1.0 MB tokens | R 5.5059 | R 16.5177 | Healthy |
Specs →
|
|
O
openai/gpt-4o
OpenAI flagship omnichannel intelligent model.
|
OpenAI | 125.0 KB tokens | R 63.8441 | R 255.3763 | Healthy |
Specs →
|
|
O
openai/gpt-4o-2024-11-20
OpenAI GPT-4o pinned snapshot release.
|
OpenAI | 125.0 KB tokens | R 63.8441 | R 255.3763 | Healthy |
Specs →
|
|
O
openai/gpt-5-pro
Advanced frontier reasoning model by OpenAI.
|
OpenAI | 390.6 KB tokens | R 383.0644 | R 3064.5155 | Healthy |
Specs →
|
|
O
openai/gpt-5-mini
Efficient high-reasoning compact GPT-5 model.
|
OpenAI | 390.6 KB tokens | R 6.3844 | R 51.0753 | Healthy |
Specs →
|
|
O
openai/gpt-6.1-terra
Massive context frontier reasoning model.
|
OpenAI | 1.0 MB tokens | R 25.5376 | R 127.6881 | Healthy |
Specs →
|
|
O
openai/gpt-6.1-terra-pro
High-precision extended context OpenAI model.
|
OpenAI | 1.0 MB tokens | R 25.5376 | R 127.6881 | Healthy |
Specs →
|
|
O
openai/gpt-6.1-sol
Speed-optimized ultra-long context model.
|
OpenAI | 1.0 MB tokens | R 51.0753 | R 255.3763 | Healthy |
Specs →
|
|
A
anthropic/claude-sonnet-5.5
Next-generation Claude Sonnet with million-token context.
|
Anthropic | 976.6 KB tokens | R 51.0753 | R 255.3763 | Healthy |
Specs →
|
|
A
anthropic/claude-sonnet-5
Claude 5 high-throughput balanced reasoning model.
|
Anthropic | 976.6 KB tokens | R 51.0753 | R 255.3763 | Healthy |
Specs →
|
|
A
anthropic/claude-sonnet-4.5
Claude Sonnet 4.5 enterprise reasoning powerhouse.
|
Anthropic | 976.6 KB tokens | R 76.6129 | R 383.0644 | Healthy |
Specs →
|
|
A
anthropic/claude-haiku-5.5
Near-instantaneous lightweight Claude execution engine.
|
Anthropic | 976.6 KB tokens | R 2.5538 | R 12.7688 | Healthy |
Specs →
|
|
A
anthropic/claude-opus-5.5
Anthropic most powerful flagship model for deep synthesis.
|
Anthropic | 976.6 KB tokens | R 102.1505 | R 510.7526 | Healthy |
Specs →
|
|
M
mistralai/mistral-large-4-0
Mistral AI top-tier multilingual flagship model.
|
Mistral AI | 1.0 MB tokens | R 17.8763 | R 53.6290 | Healthy |
Specs →
|
|
M
mistralai/mistral-medium-3-5
Balanced enterprise workhorse model by Mistral AI.
|
Mistral AI | 256.0 KB tokens | R 38.3064 | R 191.5322 | Healthy |
Specs →
|
|
M
mistralai/mistral-small-2603
Fast, cost-efficient compact reasoning model.
|
Mistral AI | 256.0 KB tokens | R 3.8306 | R 15.3226 | Healthy |
Specs →
|
|
M
mistralai/ministral-14b-2512
High performance 14B parameter edge-optimized model.
|
Mistral AI | 256.0 KB tokens | R 5.1075 | R 5.1075 | Healthy |
Specs →
|
|
M
mistralai/ministral-8b-2512
Fast lightweight reasoning model from Mistral.
|
Mistral AI | 256.0 KB tokens | R 3.8306 | R 3.8306 | Healthy |
Specs →
|
|
M
mistralai/devstral-2512
Dedicated developer coding and engineering model.
|
Mistral AI | 256.0 KB tokens | R 10.2151 | R 51.0753 | Healthy |
Specs →
|
|
G
google/gemini-4-argon
Next-generation million-token multimodal reasoning.
|
1.0 MB tokens | R 51.0753 | R 255.3763 | Healthy |
Specs →
|
|
|
G
google/gemini-3.8-flash
Low latency, high throughput flash model from Google.
|
1.0 MB tokens | R 19.1532 | R 95.7661 | Healthy |
Specs →
|
|
|
G
google/gemini-3.7-flash
Google efficient high-speed production model.
|
1.0 MB tokens | R 19.1532 | R 95.7661 | Healthy |
Specs →
|
|
|
G
google/gemini-3.6-flash
Cost-effective general intelligence model.
|
1.0 MB tokens | R 19.1532 | R 95.7661 | Healthy |
Specs →
|
|
|
G
google/gemini-3.5-flash
Stable million-token fast inference engine.
|
1.0 MB tokens | R 38.3064 | R 229.8387 | Healthy |
Specs →
|
|
|
G
google/gemini-3.5-flash-lite
Lightweight economy model for high-volume tasks.
|
1.0 MB tokens | R 7.6613 | R 63.8441 | Healthy |
Specs →
|
|
|
Q
qwen/qwen3.8-max-prime
Qwen flagship frontier model for reasoning and math.
|
Qwen | 976.6 KB tokens | R 102.1505 | R 306.4515 | Healthy |
Specs →
|
|
Q
qwen/qwen3.8-flash
Ultra-fast high performance Qwen model.
|
Qwen | 976.6 KB tokens | R 3.8306 | R 12.0027 | Healthy |
Specs →
|
|
Q
qwen/qwen3.8-27b
Efficient 27B parameter instruction model.
|
Qwen | 1000.0 KB tokens | R 3.8306 | R 47.8831 | Healthy |
Specs →
|
|
Q
qwen/qwen3.7-plus
Enhanced Qwen 3.7 reasoning and coding model.
|
Qwen | 976.6 KB tokens | R 8.1720 | R 32.6882 | Healthy |
Specs →
|
|
Q
qwen/qwen3.7-flash
Economical high-speed conversational model.
|
Qwen | 976.6 KB tokens | R 0.7661 | R 3.3199 | Healthy |
Specs →
|
|
x
x-ai/grok-4.7
xAI premier intelligence engine with direct knowledge.
|
xAI | 488.3 KB tokens | R 51.0753 | R 153.2258 | Healthy |
Specs →
|
|
x
x-ai/grok-4.6
Grok 4.6 enterprise conversational model.
|
xAI | 488.3 KB tokens | R 51.0753 | R 153.2258 | Healthy |
Specs →
|
|
x
x-ai/grok-4.5
Fast, witty, high-performance Grok release.
|
xAI | 488.3 KB tokens | R 51.0753 | R 153.2258 | Healthy |
Specs →
|
|
C
cohere/command-a-plus
Cohere flagship enterprise agentic model.
|
Cohere | 187.5 KB tokens | R 7.6613 | R 38.3064 | Healthy |
Specs →
|
|
C
cohere/command-r-plus-08-2024
Optimized for enterprise RAG and multi-step workflows.
|
Cohere | 125.0 KB tokens | R 63.8441 | R 255.3763 | Healthy |
Specs →
|
|
C
cohere/command-r-08-2024
Fast and highly capable RAG-focused model.
|
Cohere | 125.0 KB tokens | R 3.8306 | R 15.3226 | Healthy |
Specs →
|
|
M
moonshotai/kimi-k3.1-pro
Kimi million-token long context document reasoning.
|
Moonshot AI | 1.0 MB tokens | R 51.0753 | R 255.3763 | Healthy |
Specs →
|
|
M
minimax/minimax-m3.1-flash
High-speed million-token multimodal reasoning.
|
MiniMax | 1.0 MB tokens | R 3.8306 | R 15.3226 | Healthy |
Specs →
|