New Models Added: GPT-5.1-Codex-Max and Mistral Large 3
Date: 2025-12-06 / (1404-09-15)
Summary
We announce the addition of two powerful new models: GPT-5.1-Codex-Max from OpenAI, the most intelligent coding model optimized for long-horizon agentic coding tasks, and Mistral Large 3 from Mistral AI, a state-of-the-art open model with 675B total parameters and multimodal capabilities. Both models are now available through the AvalAI API.
Details
OpenAI
GPT-5.1-Codex-Max
We introduce GPT-5.1-Codex-Max (gpt-5.1-codex-max), OpenAI's most intelligent coding model purpose-built for agentic coding. This model is optimized for long-horizon, multi-step programming tasks and autonomous agent workflows. Documentation
Key Features:
- Context Window: 400,000 tokens for handling extensive codebases and documents
- Max Output Tokens: 128,000 tokens for comprehensive code generation
- Advanced Capabilities: Function calling, structured outputs, reasoning token support, vision (image input)
- Agentic Focus: Optimized for long-horizon coding tasks and autonomous programming workflows
- Knowledge Cutoff: September 30, 2024
- Endpoint Support: Available on
v1/chat/completionsandv1/responses - Streaming: Fully supported
Modalities:
- Input: Text, Image
- Output: Text
Pricing Details:
| Model | Input | Cached Input | Output |
|---|---|---|---|
| gpt-5.1-codex-max | $1.25/1M tokens | $0.125/1M tokens | $10.00/1M tokens |
Example Usage:
bash
curl https://api.avalai.ir/v1/chat/completions \
-H "Content-Type: application/json" \
-H "Authorization: Bearer $AVALAI_API_KEY" \
-d '{
"model": "gpt-5.1-codex-max",
"messages": [
{
"role": "user",
"content": "Create a complete TypeScript implementation of a distributed task queue system with worker management, task prioritization, and failure recovery."
}
]
}'python
from openai import OpenAI
client = OpenAI(api_key="your-avalai-api-key", base_url="https://api.avalai.ir/v1")
completion = client.chat.completions.create(
model="gpt-5.1-codex-max",
messages=[
{
"role": "user",
"content": "Create a complete TypeScript implementation of a distributed task queue system with worker management, task prioritization, and failure recovery.",
}
],
)
print(completion.choices[0].message.content)javascript
import { OpenAI } from "openai";
const client = new OpenAI({
apiKey: process.env.AVALAI_API_KEY,
baseURL: "https://api.avalai.ir/v1",
});
const completion = await client.chat.completions.create({
model: "gpt-5.1-codex-max",
messages: [
{
role: "user",
content: "Create a complete TypeScript implementation of a distributed task queue system with worker management, task prioritization, and failure recovery.",
},
],
});
console.log(completion.choices[0].message.content);Mistral AI
Mistral Large 3
We introduce Mistral Large 3 (mistral-large-3), Mistral AI's most capable model to date. This state-of-the-art open model is a sparse mixture-of-experts architecture with 41B active and 675B total parameters, released under the Apache 2.0 license. Documentation
Key Features:
- Architecture: Sparse Mixture-of-Experts (41B active / 675B total parameters)
- Open Source: Released under Apache 2.0 license
- Multimodal: Native image understanding capabilities
- Multilingual: Best-in-class performance on non-English/Chinese conversations with 40+ native languages
- Context Window: Large context window for extensive document processing
- LMArena Ranking: #2 in OSS non-reasoning models category (#6 overall among OSS models)
- Endpoint Support: Available on
v1/chat/completions
What Makes Mistral Large 3 Special:
- First mixture-of-experts model from Mistral since the seminal Mixtral series
- Achieves parity with best instruction-tuned open-weight models on general prompts
- Optimized checkpoint available in NVFP4 format for efficient deployment
- Can run on single 8×A100 or 8×H100 node using vLLM
Pricing Details:
| Model | Input | Cached Input | Output | Per Page |
|---|---|---|---|---|
| mistral-large-3 | $0.50/1M tokens | $0.05/1M tokens | $1.50/1M tokens | $0.001/page |
Example Usage:
bash
curl https://api.avalai.ir/v1/chat/completions \
-H "Content-Type: application/json" \
-H "Authorization: Bearer $AVALAI_API_KEY" \
-d '{
"model": "mistral-large-3",
"messages": [
{
"role": "user",
"content": "Analyze the implications of quantum computing on modern cryptography and suggest mitigation strategies."
}
]
}'python
from openai import OpenAI
client = OpenAI(api_key="your-avalai-api-key", base_url="https://api.avalai.ir/v1")
completion = client.chat.completions.create(
model="mistral-large-3",
messages=[
{
"role": "user",
"content": "Analyze the implications of quantum computing on modern cryptography and suggest mitigation strategies.",
}
],
)
print(completion.choices[0].message.content)javascript
import { OpenAI } from "openai";
const client = new OpenAI({
apiKey: process.env.AVALAI_API_KEY,
baseURL: "https://api.avalai.ir/v1",
});
const completion = await client.chat.completions.create({
model: "mistral-large-3",
messages: [
{
role: "user",
content: "Analyze the implications of quantum computing on modern cryptography and suggest mitigation strategies.",
},
],
});
console.log(completion.choices[0].message.content);Using Mistral SDK:
python
from mistralai import Mistral
client = Mistral(server_url="https://api.avalai.ir", api_key="avalai-api-key")
response = client.chat.complete(
model="mistral-large-3",
messages=[
{
"role": "user",
"content": "Explain the benefits of mixture-of-experts architecture in large language models.",
}
],
)
print(response.choices[0].message.content)javascript
import { Mistral } from "mistralai";
const client = new Mistral({
apiKey: "[REDACTED]",
baseURL: "https://api.avalai.ir",
});
const response = await client.chat.complete({
model: "mistral-large-3",
messages: [
{
role: "user",
content: "Explain the benefits of mixture-of-experts architecture in large language models.",
},
],
});
console.log(response.choices[0].message.content);