Vikasit Titan 1.6T
Flagship frontier MoE. 1.6T parameters, 1M-token context. Most capable model.
Overview
Vikasit Titan 1.6T is the flagship frontier model — 1.6T total / 49B active, trained on 32T+ tokens, with a 1M-token context window and state-of-the-art coding, reasoning, and agentic performance. The most capable model in the Vikasit lineup. Served live via the Vikasit API.
Specifications
- Total parameters
- 1.6T total
- Active parameters
- 49B active
- Architecture
- Mixture-of-Experts — Compressed Sparse Attention (CSA) + Heavily Compressed Attention (HCA), mHC connections
- Context window
- 1M tokens
- Modalities
- Text in → text out
- License
- MIT
Capabilities
- State-of-the-art coding and reasoning
- 1M-token context window
- Frontier agentic / terminal performance
- Novel compressed-attention efficiency
Benchmarks
| Benchmark | Score |
|---|---|
| MMLU-Pro | 87.5 |
| GPQA-Diamond | 90.1 |
| LiveCodeBench | 93.5 |
| SWE-bench Verified | 80.6 |
| Humanity's Last Exam | 37.7 |
| Terminal-Bench 2.0 | 67.9 |
| HMMT Feb 2026 | 95.2 |
Numbers from the DeepSeek V4-Pro official model card + technical report (V4-Pro Max flagship column). Note: the base is text-only despite its scale; context is 1M tokens, text.
Hardware & deployment
| Precision | Memory |
|---|---|
| bf16 | ~3.2 TB |
| INT4 | ~800 GB |
Quick start
Call Vikasit Titan 1.6T through the OpenAI-compatible Vikasit AI API at https://api.vikasit.ai/v1 using the model id vikasit-titan-1.6t.
# pip install openai
import os
from openai import OpenAI
client = OpenAI(
base_url="https://api.vikasit.ai/v1",
api_key=os.environ["VIKASIT_API_KEY"],
)
resp = client.chat.completions.create(
model="vikasit-titan-1.6t",
messages=[
{"role": "user", "content": "Explain Vikasit Titan 1.6T in one sentence."}
],
)
print(resp.choices[0].message.content)# or with curl
curl https://api.vikasit.ai/v1/chat/completions \
-H "Authorization: Bearer $VIKASIT_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "vikasit-titan-1.6t",
"messages": [{"role": "user", "content": "Hello"}]
}'Limitations
- Text-only (no vision/audio)
- Frontier-scale serving — cluster required
Vikasit Titan 1.6T FAQ
How much does Vikasit Titan 1.6T cost?
Vikasit Titan 1.6T is served through the Vikasit AI API on usage-based, pay-as-you-go pricing billed per million input and output tokens — see the Vikasit AI pricing page for current rates. Because it is built on the open-weight DeepSeek V4-Pro (DeepSeek, MIT), you can also self-host the weights for free under the MIT licence and pay only for your own compute.
Is Vikasit Titan 1.6T open weight?
Yes. Vikasit Titan 1.6T is built on DeepSeek V4-Pro (DeepSeek, MIT) and distributed under the MIT licence, so the weights are openly available for self-hosting, fine-tuning, and commercial use, subject to the upstream licence terms.
How do I use Vikasit Titan 1.6T with the OpenAI SDK?
The Vikasit AI API is OpenAI-compatible. Point the OpenAI client's base URL at https://api.vikasit.ai/v1, set your Vikasit API key, and pass "vikasit-titan-1.6t" as the model. The quick-start snippet above shows the exact Python call.
What context window does Vikasit Titan 1.6T support?
Vikasit Titan 1.6T supports a 1M tokens context window. It is a 1.6T total (49B active) Mixture-of-Experts — Compressed Sparse Attention (CSA) + Heavily Compressed Attention (HCA), mHC connections model — full specifications are listed in the table above.
License & attribution
MIT
Built on DeepSeek V4-Pro (DeepSeek, MIT). Upstream copyright, license, and attribution notices are retained.