Skip to main content
Use these models with the chat completions endpoint. Reference a model by its id in the model field.
The exact set your key can call is returned by GET /v1/models. For per-model rates, see deepshi.ai.

​
Deepshi models

All of Deepshi’s own models are uncensored.

Deepshi 2.0

deepshi-2.0 · Context 128KHighly uncensored, with open, minimally-filtered responses. The most direct option. Tools   Reasoning   Vision

Deepshi 3.0

deepshi-3.0 · Context 256KMultimodal flagship with strong visual understanding and agentic coding. Tools   Reasoning   Vision

​
Third-party models

The latest models from leading providers, through the same endpoint, key, and balance. Recently added models appear first.

DeepSeek V4.1 Flash

deepseek-v4.1-flash · Context 1MDeepSeek’s efficient open-weight multimodal MoE for coding, long-running agents, and analysis across very large inputs. Tools   Reasoning   Vision

Ling 3.0 Flash VL

ling-3.0-flash-vl · Context 131KInclusionAI’s fast open-weight multimodal MoE for image and video understanding, tool use, and visual-agent workflows. Tools   Reasoning   Vision

GPT-6 Astra

gpt-6-astra · Context 1MOpenAI’s GPT-6 flagship for complex reasoning, coding, and multi-step agentic work. Reasoning is always on and cannot be turned off. Tools   Reasoning   Vision

Gemini 3.8 Flash

gemini-3.8-flash · Context 1MGoogle’s latest fast multimodal model, built for agentic and multi-step reasoning work where latency matters. Tools   Reasoning   Vision

Claude Fable 5.1

claude-fable-5.1 · Context 1MAnthropic’s updated agentic flagship, refining Fable’s long-running coding and research with minimal supervision. Tools   Reasoning   Vision

DeepSeek V4 Flash Vision Exp

deepseek-v4-flash-vision-exp · Context 1MDeepSeek’s experimental vision-enabled Flash variant, bringing the fast, low-cost line to image understanding. Tools   Reasoning   Vision

Nemotron 3.5 Lightning

nemotron-3.5-lightning · Context 256KNVIDIA’s compact open MoE, activating 3B of 30B parameters for high-throughput agentic work. Tools   Reasoning

Grok 4.6

grok-4.6 · Context 500KxAI’s newest frontier model, succeeding Grok 4.5 on coding, knowledge work, and STEM. Tools   Reasoning   Vision

GLM 5.3

glm-5.3 · Context 1MZ.ai’s flagship reasoning model for project-level engineering and sustained agent runs. Tools   Reasoning

GLM 5.3 Flash

glm-5.3-flash · Context 1.3MZ.ai’s efficient natively multimodal model, a low-cost pick for coding and long agent tasks. Tools   Reasoning   Vision

Qwen3.8 Flash

qwen3.8-flash · Context 1MAlibaba’s fast, low-cost multimodal model for coding, agents, and visual understanding. Tools   Reasoning   Vision

Muse Glimmer 30B

muse-glimmer-30b · Context 128KMeta’s dense 30B open-weight multimodal model, distilled from Muse Spark for on-device-class agents. Tools   Reasoning   Vision

DeepSeek V4 Pro 0813

deepseek-v4-pro-0813 · Context 1MThe 0813 general-availability release of DeepSeek’s large MoE for advanced reasoning and coding. Tools   Reasoning

Qwen3.8 2.4T A95B

qwen3.8-2.4t-a95b · Context 1MThe open-weight MoE counterpart to Qwen3.8 Max, with 95B active of 2.4T total parameters. Tools   Reasoning

Gemini 3.7 Flash

gemini-3.7-flash · Context 1MGoogle’s latest Flash-class multimodal model, with text, image, audio, video, and file input. Tools   Reasoning   Vision

Qwen3.8 27B

qwen3.8-27b · Context 256KA dense 27B open-weight Qwen model with multimodal input for coding and professional work. Tools   Reasoning   Vision

LongCat 2.0

longcat-2.0 · Context 1MMeituan’s sparse MoE with 48B active of 1.6T parameters, aimed at repository-scale coding. Tools   Reasoning

Qwen3.8 Max

qwen3.8-max · Context 1MAlibaba’s flagship Qwen3.8, a reasoning model for complex problem solving and agentic workflows. Tools   Reasoning   Vision

DeepSeek V4 Flash 0731

deepseek-v4-flash-0731 · Context 1MAn updated post-training revision of DeepSeek’s efficiency-focused MoE, activating 13B of 284B. Tools   Reasoning

Inkling Small

inkling-small · Context 512KThe compact Inkling, a multimodal MoE activating 12B of 276B parameters, with image and audio input. Tools   Reasoning   Vision

Laguna S 2.1

laguna-s-2.1 · Context 1MPoolside’s coding agent model, a 118B MoE with 8B active parameters. Tools   Reasoning

Gemini 3.5 Flash Lite

gemini-3.5-flash-lite · Context 1MGoogle’s lightest Flash-class model, built for fast, high-volume multimodal work. Tools   Reasoning   Vision

Gemini 3.6 Flash

gemini-3.6-flash · Context 1MGoogle’s high-efficiency multimodal model for coding and agents, with audio and video input. Tools   Reasoning   Vision

Claude Opus 5

claude-opus-5 · Context 1MAnthropic’s flagship for demanding reasoning and end-to-end software tasks. Tools   Reasoning   Vision

Inkling

inkling · Context 1MThinking Machines’ open-weight MoE model for reasoning, coding, and agentic work, with image and audio input. Tools   Reasoning   Vision

Kimi K3

kimi-k3 · Context 1MMoonshot’s open-weight reasoning model for large codebases, tool use, and long autonomous runs. Tools   Reasoning   Vision

Claude Sonnet 5

claude-sonnet-5 · Context 1MAnthropic’s most capable Sonnet, balancing speed and depth for coding and agent workflows. Tools   Reasoning   Vision

Claude Fable 5

claude-fable-5 · Context 1MAnthropic’s agentic Claude 5 model for long-running, minimally-supervised coding and research. Tools   Reasoning   Vision

GPT-5.6 Sol

gpt-5.6-sol · Context 1MThe flagship of OpenAI’s GPT-5.6 series, built for complex reasoning and multi-step coding. Tools   Reasoning   Vision

GPT-5.6 Terra

gpt-5.6-terra · Context 1MThe balanced GPT-5.6 tier for everyday coding, reasoning, and agentic tasks. Tools   Reasoning   Vision

GPT-5.6 Luna

gpt-5.6-luna · Context 1MThe fast, cost-efficient GPT-5.6 tier for high-volume, latency-sensitive work. Tools   Reasoning   Vision

Grok 4.5

grok-4.5 · Context 500KxAI’s most capable model, with frontier performance across coding, knowledge work, and STEM. Tools   Reasoning   Vision

Kimi K2.7 Code

kimi-k2.7-code · Context 256KMoonshot’s coding-focused Kimi model for end-to-end programming over long contexts. Tools   Reasoning   Vision

Nemotron 3 Ultra

nemotron-3-ultra-550b-a55b · Context 1MNVIDIA’s open frontier reasoning and orchestration model with a 1M context. Tools   Reasoning

MiniMax M3

minimax-m3 · Context 1MMiniMax multimodal model for long-horizon agentic work, with a 1M context. Tools   Reasoning   Vision

Step 3.7 Flash

step-3.7-flash · Context 256KStepFun’s efficient multimodal model with native image and video understanding. Tools   Reasoning   Vision

Grok Build 0.1

grok-build-0.1 · Context 256KxAI’s fast coding model built for agentic software engineering. Tools   Reasoning   Vision

Granite 4.1 8B

granite-4.1-8b · Context 128KIBM’s dense 8B model for enterprise retrieval, tools, and structured text. Tools

Qwen3.6 35B A3B

qwen3.6-35b-a3b · Context 256KAlibaba’s open multimodal MoE balancing quality and efficient inference. Tools   Reasoning   Vision

Qwen3.6 27B

qwen3.6-27b · Context 256KDense 27B Qwen model with multimodal input for general-purpose work. Tools   Reasoning   Vision

DeepSeek V4 Pro

deepseek-v4-pro · Context 1MDeepSeek’s large MoE for advanced reasoning and coding over a 1M context. Tools   Reasoning

DeepSeek V4 Flash

deepseek-v4-flash · Context 1MEfficiency-tuned DeepSeek MoE for fast, low-cost reasoning at a 1M context. Tools   Reasoning

MiMo V2.5 Pro

mimo-v2.5-pro · Context 1MXiaomi’s flagship for agentic capability and complex software engineering. Tools   Reasoning

GPT-5.4 Mini

gpt-5.4-mini · Context 400KFaster, efficient GPT-5.4 variant for high-throughput workloads. Tools   Reasoning   Vision

GPT-5.4 Nano

gpt-5.4-nano · Context 400KThe lightest GPT-5.4 model, tuned for low-latency, high-volume use. Tools   Reasoning   Vision

GLM 5 Turbo

glm-5-turbo · Context 256KZ.ai’s fast-inference model tuned for agent-driven workflows. Tools   Reasoning

Nemotron 3 Super

nemotron-3-super-120b-a12b · Context 1MNVIDIA’s 120B hybrid MoE for efficient multi-agent applications. Tools   Reasoning

MiniMax M2.5

minimax-m2.5 · Context 200KMiniMax model trained for real-world productivity and coding. Tools   Reasoning

Qwen3 Coder Next

qwen3-coder-next · Context 256KQwen’s open coding model for coding agents and local development. Tools

GLM 4.7 Flash

glm-4.7-flash · Context 200KZ.ai’s 30B-class model balancing performance with agentic coding. Tools   Reasoning

Gemma 4 26B A4B

gemma-4-26b-a4b-it · Context 256KGoogle’s instruction-tuned MoE delivering near-31B quality at lower cost. Tools   Reasoning   Vision

MiniMax M2.7

minimax-m2.7 · Context 200KNext-generation MiniMax model for autonomous, agentic productivity. Tools   Reasoning

GPT-OSS 120B

gpt-oss-120b · Context 128KOpenAI’s open-weight 117B MoE for high-reasoning, agentic use. Tools   Reasoning

GPT-OSS 20B

gpt-oss-20b · Context 128KOpenAI’s open-weight 21B MoE under Apache 2.0, tuned for efficient reasoning. Tools   Reasoning

Claude Opus 4.8

claude-opus-4.8 · Context 1MAnthropic’s most capable Opus model, with deep reasoning over long, complex tasks. Tools   Reasoning   Vision

Claude Opus 4.7

claude-opus-4.7 · Context 1MOpus-family model built for long-running, asynchronous agents. Tools   Reasoning   Vision

Claude Sonnet 4.6

claude-sonnet-4.6 · Context 1MStrong all-rounder for coding, agents, and professional work. Tools   Reasoning   Vision

GPT-5.5

gpt-5.5 · Context 1MOpenAI’s frontier model for complex professional workloads. Tools   Reasoning   Vision

GPT-5.4

gpt-5.4 · Context 1MFrontier model unifying the GPT and Codex lines, strong at agentic coding. Tools   Reasoning   Vision

GPT-4.1

gpt-4.1 · Context 1MTuned for precise instruction-following and software engineering. Tools   Vision

GPT-4o

gpt-4o · Context 128KGPT-4-class intelligence that runs faster and cheaper. Solid general-purpose pick. Tools   Vision

Gemini 3.5 Flash

gemini-3.5-flash · Context 1MGoogle’s high-efficiency model with near-Pro reasoning at Flash speed and cost. Tools   Reasoning   Vision

Grok 4.3

grok-4.3 · Context 1MxAI reasoning model suited to agentic workflows and high factual accuracy. Tools   Reasoning   Vision

Grok 4.20

grok-4.20 · Context 2MFast xAI reasoning model with strong tool-calling and low hallucination. Tools   Reasoning   Vision

GLM-5.2

glm-5.2 · Context 1MLarge-scale reasoning model for long-horizon agent and engineering work. Tools   Reasoning

GLM-5.1

glm-5.1 · Context 200KA major step up in coding ability on long-horizon tasks. Tools   Reasoning

GLM-5

glm-5 · Context 200KZ.ai’s flagship open model for complex systems and agent workflows. Tools   Reasoning

Kimi K2.6

kimi-k2.6 · Context 256KBuilt for long-horizon coding, UI/UX generation, and multi-agent orchestration. Tools   Reasoning   Vision

Gemma 4 31B

gemma-4-31b-it · Context 256KA ~31B dense open model with an optional reasoning mode and native tools. Tools   Reasoning   Vision
Not sure where to start? Use deepshi-2.0 for highly uncensored responses, or deepshi-3.0 for an all-round flagship with tools and reasoning.