Tencent MoE language model (295B parameters, 21B active) with a 256K context window, strong at agentic tasks, reasoning and coding.
Context window
256K
tokens
Parameters
295B (21B aktywnych)
parameters
Access:DownloadDeployment:☁ Cloud💻 Local
Overview
Access & deployment
Download
CloudLocal
Weights: Open weights
Key parameters
📏 Context: 256K
🧩 Parameters: 295B (21B aktywnych)
✓ Tools · ✓ Fine-tuning
📥 Input: text
Technical specification
Context window
256K
tokens
Parameters
295B (21B aktywnych)
parameters
License
Apache 2.0
Hardware requirements
The 295B MoE model (21B active) in BF16 requires around 8 high-memory GPUs (e.g. H100/H200); an FP8 variant lowers requirements. Recommended inference engines: vLLM or SGLang.
Features:✓ Tool use✓ Fine-tuning
Modalities
⬇ Input
text
⬆ Output
textcode
Capabilities and applications
Native model capabilities
Coding
Generating, analysing and modifying code in many programming languages. Covers writing functions, debugging, refactoring, code review, and creating tests. Measured by benchmarks such as HumanEval and SWE-bench.
Category: coding
Agentic coding
Multi-hour, multi-step programming tasks performed autonomously by the model: cloning a repository, running tests, iterating on fixes, integrating with CLI tools. Characteristic of Codex variants (GPT-5.1-Codex-Mini, Codex-Max).
Category: coding
Agentic capability
The model's ability to autonomously plan and execute multi-step tasks by sequentially using tools, maintaining context, and adapting to intermediate results.
Category: planning
Tool use
The model's ability to call external functions, APIs and tools during a conversation: calculator, search engine, code editor, database. The model decides when and how to use a tool and interprets its result.
Category: planning
Advanced reasoning
The ability to perform multi-step, structured reasoning: analysing problems, planning steps, and drawing conclusions from hypotheses. Reasoning-first models (e.g. GPT-5.1 Thinking) dedicate a portion of inference to chains of thought before responding.
Category: reasoning
Long context
Support for large context windows — tens to hundreds of thousands (or millions) of input tokens. Enables analysis of entire codebases, long documents, and many parallel conversations without losing earlier information. GPT-5.1 supports 400,000 tokens.
Category: language
Mathematical reasoning
The model's ability to solve mathematical tasks requiring multi-step reasoning — equations, proofs, combinatorics, geometry, calculus and competition-level problems.
Category: reasoning
Financial modeling
Building financial models: DCF, company valuations, scenario analysis, budget forecasts, P&L sheets, cap table models. Requires precise numerical reasoning and knowledge of accounting conventions.
Category: reasoning
Multilingual
Competence in many natural languages (from a few to over a hundred): understanding, generation, translation, and code-switching within a single conversation. Frontier models support a wide range of languages with comparable quality.
Category: language
Benchmark results
3 benchmarks
GPQA
accuracy · GPQA Diamond
90.4%
📄 Oficjalna karta modelu (Hugging Face)
SWE-Bench Pro
resolved · SWE-Bench Pro
57.9%
📄 Oficjalna karta modelu (Hugging Face)
SWE-bench
resolved · SWE-bench Multilingual
75.8%
📄 Oficjalna karta modelu (Hugging Face)
Technical architecture
Core Architecture
Model Form