Qwen3-235B: the model previously served by Gonka

Qwen3-235B-A22B-Instruct-2507-FP8 is a Large Language Model (LLM) from the Qwen3 family, developed by the Qwen team at Alibaba Cloud. The full name deciphers as: Qwen3 — the third generation of the series, 235B — 235 billion parameters in total, A22B — 22 billion active parameters per query, Instruct — a version trained to follow instructions, 2507 — July 2025 release, FP8 — 8-bit quantization for memory optimization.

A key architectural feature is MoE (Mixture of Experts). Unlike "dense" models (GPT-5.5, Claude Sonnet 4.6), where each token passes through all parameters, an MoE model activates only a subset of "experts" — specialized blocks of the neural network — for each query. In the case of Qwen3-235B, out of 235 billion parameters, only 22 billion are activated for each token — less than 10%. This delivers quality comparable to models with 200B+ parameters at the computational cost of a 22B model.

Practically, this means: the model is smarter than one might expect from its speed. It processes queries significantly faster than dense models of comparable quality, while requiring orders of magnitude less VRAM for inference. This is why MoE has become the dominant architecture for the largest models of 2025–2026.

The context window for Qwen3-235B is 131,072 tokens (~100,000 words) — sufficient for analyzing entire books, codebases, or long legal documents in a single query. The model supports 119 languages, including Russian, English, Chinese, Arabic, Hindi, and dozens of others — making it one of the most multilingual models on the market.

Qwen3-235B competes with the largest closed and open models. Here is a comparison of key characteristics:

Qwen3-235B demonstrates a quality level comparable to GPT-5.5 and Claude Sonnet 4.6 on most benchmarks, while as an open-weights MoE model, it is significantly cheaper than proprietary analogs: the MoE architecture activates only a portion of parameters per request, drastically reducing computational costs. The same principle of cheap decentralized inference is applied by the Gonka network to its current models — Kimi K2.6 and MiniMax M2.7, available via JoinGonka Gateway for $0.0047 per 1M tokens (hundreds to thousands of times cheaper than GPT-5.5 and Claude).

On MMLU-Pro, HumanEval, MATH-500, and GSM8K benchmarks, the model is among the top three open-source models, trailing only DeepSeek-R1 in mathematical reasoning tasks. In code generation, translation, and instruction-following tasks, Qwen3-235B consistently outperforms Llama 4 Maverick and is comparable to Claude Sonnet 4.6.

When Qwen3-235B was the primary model of the network, it operated on the Gonka network in a distributed manner — via the DiLoCo protocol, adapted for inference. The full model in FP8 format requires about 640 GB of VRAM (VRAM), which is impossible to fit on a single GPU — even an H100 80GB or H200 141GB is not enough. Therefore, the model was split by layers (tensor parallelism + pipeline parallelism) across multiple MLNodes.

In practice, Qwen3-235B ran on a cluster of 8—16 GPU-nodes, each with at least 40 GB of VRAM. Transfer Agents routed the request to the required cluster, vLLM on each node processed its fragment of the model, and the results were aggregated and returned to the user. The entire process took hundreds of milliseconds — the user did not feel that their request was processed by a dozen GPUs at different points on the planet. The network applies the same principle of distributed inference today — to the currently active models.

An important technical detail: Gonka uses vLLM as the serving engine. vLLM is an open-source project that provides high-performance text generation via PagedAttention — an algorithm that optimizes VRAM usage when processing multiple requests in parallel. This allows the network to serve thousands of concurrent users without quality degradation.

The model supports native tool calling — invoking functions and tools directly from the model's response. This capability was added to Gonka via PR #767 with a threshold of 0.958 for determining tool calls. On Qwen3-235B, this allowed developers to build AI agents that interact with external APIs, databases, and tools — all through a single request. Tool calling support has been maintained in the network's current models.

The Gonka network counts over 4,000 GPUs (H100, H200, A100, RTX 4090, and others) combined into 120+ MLNodes. It is one of the largest distributed GPU networks for AI inference in the world — and if previously this power was directed at Qwen3-235B, today it serves the network's active models — Kimi K2.6, MiniMax M2.7, and DeepSeek V4 Flash.

Direct answer: no, not via the Gonka network anymore. Qwen3-235B has been removed from the network, so you can no longer call it via the qwen3-235b-a22b identifier through our Gateway. Since the model is open-weights (Apache 2.0), you can still run it yourself or access it via third-party open-weights hosting providers—for example, OpenRouter (approximately $0.071/$0.100 per 1M tokens).

If you are looking for the same cheap decentralized inference that brings users to Gonka, it is still there; it just runs on current network models now. Via the JoinGonka API Gateway, you can access Kimi K2.6, MiniMax M2.7, and DeepSeek V4 Flash via an OpenAI-compatible API: any code written for OpenAI works without changes—simply replace the URL, API key, and model name.

Example request to an active model:

curl https://gate.joingonka.ai/api/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "moonshotai/Kimi-K2.6",
    "messages": [{"role": "user", "content": "Explain the MoE architecture"}]
  }'

Cost: $0.0047 per 1 million tokens — this is ~800 times cheaper than GPT-5.5 ($5.00/1M) and ~500 times cheaper than Claude Sonnet 4.6 ($3.00/1M). Upon registration, you receive free 1.5M tokens for testing.

The Gateway is compatible with popular development tools: Quick Start describes connecting via Python, Node.js, and curl. IDE integrations—Cursor, Continue, Cline, Aider, and Claude Code—as well as AI agent frameworks like LangChain, n8n, LibreChat, and Open WebUI are also supported.

For a quick start:

1. Register at gate.joingonka.ai (connect a wallet or create a new one)
2. Get an API key in the Dashboard
3. Replace api.openai.com with gate.joingonka.ai/api in your code
4. Specify an active network model — moonshotai/Kimi-K2.6, MiniMaxAI/MiniMax-M2.7, or deepseek-ai/DeepSeek-V4-Flash-0731 (the full list is in the GET /v1/models endpoint)

Enterprise-grade decentralized inference at hobby-project prices has not gone anywhere—Qwen3-235B simply gave way to newer network models. The same price-to-quality ratio now works with Kimi K2.6, MiniMax M2.7, and DeepSeek V4 Flash.