Skip to content

A reactive notebook for Python — run reproducible experiments, query with SQL, execute as a script, deploy as an app, and version with git. Stored as pure Python. All in a modern, AI-native editor.

Update LLM model catalog (#11002)

github-actions[bot]e32a63b26aa7
diff --git a/packages/llm-info/data/models.yml b/packages/llm-info/data/models.yml
index 63199e7ddf..476ba14a85 100644
--- a/packages/llm-info/data/models.yml
+++ b/packages/llm-info/data/models.yml
@@ -1,8 +1,18 @@
anthropic:
+ - name: Claude Opus 5.5
+ model: claude-opus-5-5
+ description: "Anthropic's strongest Opus model for coding, agents, and knowledge work"
+ roles: [chat, edit]
+ capabilities: [thinking, tool_calling]
+ input_types: [text, image, pdf]
+ output_types: [text]
+ release_date: 2026-09-22
+ cost: {input: 4, output: 20}
+
- name: Claude Fable 5.1
model: claude-fable-5-1
- description: "Anthropic's most capable generally available model for ambitious, long-running coding, agentic, research, and enterprise workflows"
+ description: "Anthropic model for ambitious, long-running coding, agentic, research, and enterprise workflows"
roles: [chat, edit]
capabilities: [thinking, tool_calling]
input_types: [text, image, pdf]
@@ -50,16 +60,6 @@ anthropic:
release_date: 2026-05-28
cost: {input: 5, output: 25}
- - name: Claude Opus 4.7
- model: claude-opus-4-7
- description: "Anthropic Opus model built for long-running, asynchronous agents"
- roles: [chat, edit]
- capabilities: [thinking, tool_calling]
- input_types: [text, image, pdf]
- output_types: [text]
- release_date: 2026-04-16
- cost: {input: 5, output: 25}
-
- name: Claude Sonnet 4.6
model: claude-sonnet-4-6
description: "Anthropic Sonnet model for coding, agents, and professional work"
@@ -70,47 +70,57 @@ anthropic:
release_date: 2026-02-17
cost: {input: 3, output: 15}
- - name: Claude Opus 4.6
- model: claude-opus-4-6
- description: "Anthropic Opus model for coding and long-running professional tasks"
+ - name: Claude Haiku 4.5
+ model: claude-haiku-4-5
+ description: "Anthropic's fastest and most efficient model, delivering near-frontier intelligence at a fraction of the cost and latency of larger Claude models"
roles: [chat, edit]
capabilities: [thinking, tool_calling]
input_types: [text, image, pdf]
output_types: [text]
- release_date: 2026-02-05
- cost: {input: 5, output: 25}
+ release_date: 2025-10-15
+ cost: {input: 1, output: 5}
- - name: Claude Opus 4.5
- model: claude-opus-4-5
- description: "Anthropic Opus reasoning model for complex software engineering, agentic workflows, and long-horizon computer use"
+openai:
+ - name: GPT-6 Luna
+ model: gpt-6-luna
+ description: "OpenAI's most efficient model for focused, high-volume tasks"
roles: [chat, edit]
capabilities: [thinking, tool_calling]
input_types: [text, image, pdf]
output_types: [text]
- release_date: 2025-11-24
- cost: {input: 5, output: 25}
+ release_date: 2026-09-22
+ cost: {input: 0.1, output: 0.5}
- - name: Claude Haiku 4.5
- model: claude-haiku-4-5
- description: "Anthropic's fastest and most efficient model, delivering near-frontier intelligence at a fraction of the cost and latency of larger Claude models"
+ - name: GPT-6 Sol
+ model: gpt-6-sol
+ description: "OpenAI reasoning model for demanding tasks that balances capability, latency, and cost"
roles: [chat, edit]
capabilities: [thinking, tool_calling]
input_types: [text, image, pdf]
output_types: [text]
- release_date: 2025-10-15
- cost: {input: 1, output: 5}
+ release_date: 2026-09-22
+ cost: {input: 2, output: 10}
- - name: Claude Sonnet 4.5
- model: claude-sonnet-4-5
- description: "Anthropic Sonnet model optimized for real-world agents and coding workflows"
+ - name: Daybreak Blue
+ model: gpt-daybreak-blue-latest
+ description: "OpenAI model alias for authorized defensive cybersecurity work with reduced refusals"
roles: [chat, edit]
capabilities: [thinking, tool_calling]
input_types: [text, image, pdf]
output_types: [text]
- release_date: 2025-09-29
- cost: {input: 3, output: 15}
+ release_date: 2026-08-07
+ cost: {input: 4, output: 20}
+
+ - name: Daybreak Red
+ model: gpt-daybreak-red-latest
+ description: "OpenAI model alias for advanced authorized cybersecurity work, including penetration testing and exploit validation"
+ roles: [chat, edit]
+ capabilities: [thinking, tool_calling]
+ input_types: [text, image]
+ output_types: [text]
+ release_date: 2026-08-07
+ cost: {input: 12.5, output: 75}
-openai:
- name: GPT-6 Astra
model: gpt-6-astra
description: "OpenAI model for advanced analysis, software engineering, deep research, scientific work, and document creation"
@@ -171,56 +181,17 @@ openai:
release_date: 2026-04-23
cost: {input: 5, output: 30}
- - name: GPT-5.4 mini
- model: gpt-5.4-mini
- description: "GPT-5.4 mini brings the core capabilities of GPT-5.4 to a faster, more efficient model optimized for high-throughput workloads"
- roles: [chat, edit]
- capabilities: [thinking, tool_calling]
- input_types: [text, image]
- output_types: [text]
- release_date: 2026-03-17
- cost: {input: 0.75, output: 4.5}
-
- - name: GPT-5.4 nano
- model: gpt-5.4-nano
- description: "Most lightweight and cost-efficient variant of the GPT-5.4 family, optimized for speed-critical and high-volume tasks"
- roles: [chat, edit]
- capabilities: [thinking, tool_calling]
- input_types: [text, image]
- output_types: [text]
- release_date: 2026-03-17
- cost: {input: 0.2, output: 1.25}
-
- - name: GPT-5.4
- model: gpt-5.4
- description: "OpenAI model that unified the Codex and GPT lines into a single system"
- roles: [chat, edit]
- capabilities: [thinking, tool_calling]
- input_types: [text, image, pdf]
- output_types: [text]
- release_date: 2026-03-05
- cost: {input: 2.5, output: 15}
-
- - name: GPT-5.3 Codex
- model: gpt-5.3-codex
- description: "Agentic coding model that combines GPT-5.2-Codex software engineering performance with GPT-5.2 reasoning and professional knowledge"
+google:
+ - name: Claude Opus 5.5
+ model: claude-opus-5-5@default
+ description: "Anthropic's strongest Opus model for coding, agents, and knowledge work, available through Vertex AI only"
roles: [chat, edit]
capabilities: [thinking, tool_calling]
input_types: [text, image, pdf]
output_types: [text]
- release_date: 2026-02-05
- cost: {input: 1.75, output: 14}
-
- - name: ChatGPT Image
- model: chatgpt-image-latest
- description: "Previous-generation image model used in ChatGPT for image generation and editing from text and image inputs"
- roles: [chat, edit]
- capabilities: []
- input_types: [text, image]
- output_types: [text, image]
- release_date: 2025-12-16
+ release_date: 2026-09-22
+ cost: {input: 4, output: 20}
-google:
- name: Grok 4.6
model: xai/grok-4.6
description: "xAI frontier model for coding, knowledge work, and STEM"
@@ -243,7 +214,7 @@ google:
- name: Claude Fable 5.1
model: claude-fable-5-1@default
- description: "Anthropic's most capable generally available model for ambitious, long-running coding, agentic, research, and enterprise workflows, available through Vertex AI only"
+ description: "Anthropic model for ambitious, long-running coding, agentic, research, and enterprise workflows, available through Vertex AI only"
roles: [chat, edit]
capabilities: [thinking, tool_calling]
input_types: [text, image, pdf]
@@ -271,46 +242,6 @@ google:
release_date: 2026-08-13
cost: {input: 0.75, output: 3.75}
- - name: Gemini 3.6 Flash
- model: gemini-3.6-flash
- description: "Google's high-efficiency model for coding, agentic workflows, and web and app development"
- roles: [chat, edit]
- capabilities: [thinking, tool_calling]
- input_types: [text, image, pdf]
- output_types: [text]
- release_date: 2026-07-21
- cost: {input: 1.5, output: 7.5}
-
- - name: Gemini 3.5 Flash Lite
- model: gemini-3.5-flash-lite
- description: "Google's high-efficiency model with upgraded agentic capabilities, suited for focused subagent tasks in multi-agent workflows"
- roles: [chat, edit]
- capabilities: [thinking, tool_calling]
- input_types: [text, image, pdf]
- output_types: [text]
- release_date: 2026-07-21
- cost: {input: 0.3, output: 2.5}
-
- - name: Gemini 3.5 Flash
- model: gemini-3.5-flash
- description: "Google's high-efficiency multimodal model, bringing near-Pro level coding and reasoning at Flash-tier cost and speed"
- roles: [chat, edit]
- capabilities: [thinking, tool_calling]
- input_types: [text, image, pdf]
- output_types: [text]
- release_date: 2026-05-19
- cost: {input: 1.5, output: 9}
-
- - name: Gemini 3.1 Flash Lite
- model: gemini-3.1-flash-lite
- description: "Google's GA high-efficiency multimodal model optimized for low-latency, high-volume workloads"
- roles: [chat, edit]
- capabilities: [thinking, tool_calling]
- input_types: [text, image, pdf]
- output_types: [text]
- release_date: 2026-05-07
- cost: {input: 0.25, output: 1.5}
-
- name: Gemma 4 26B A4B IT
model: gemma-4-26b-a4b-it
description: "Instruction-tuned Mixture-of-Experts (MoE) model from Google DeepMind"
@@ -349,26 +280,6 @@ google:
release_date: 2026-05-28
cost: {input: 0.5, output: 60}
- - name: Gemini 3.1 Pro Preview
- model: gemini-3.1-pro-preview
- description: "Google's frontier reasoning model, delivering enhanced software engineering performance, improved agentic reliability, and more efficient token usage across complex workflows"
- roles: [chat, edit]
- capabilities: [thinking, tool_calling]
- input_types: [text, image, pdf]
- output_types: [text]
- release_date: 2026-02-19
- cost: {input: 2, output: 12}
-
- - name: Gemini 3.1 Pro Preview Custom Tools
- model: gemini-3.1-pro-preview-customtools
- description: "Variant of Gemini 3.1 Pro with refined tool selection that avoids overusing a generic bash tool"
- roles: [chat, edit]
- capabilities: [thinking, tool_calling]
- input_types: [text, image, pdf]
- output_types: [text]
- release_date: 2026-02-19
- cost: {input: 2, output: 12}
-
- name: Claude Opus 5
model: claude-opus-5@default
description: "Anthropic Claude Opus 5 on Google Vertex AI only (not available via the Gemini API)"
@@ -400,6 +311,26 @@ google:
cost: {input: 5, output: 25}
bedrock:
+ - name: Claude Opus 5.5 (Global)
+ model: global.anthropic.claude-opus-5-5
+ description: "Anthropic's strongest Opus model for coding, agents, and knowledge work"
+ roles: [chat, edit]
+ capabilities: [thinking, tool_calling]
+ input_types: [text, image, pdf]
+ output_types: [text]
+ release_date: 2026-09-22
+ cost: {input: 4, output: 20}
+
+ - name: GPT-6 Luna (Global)
+ model: global.openai.gpt-6-luna
+ description: "OpenAI's most efficient model for focused, high-volume tasks"
+ roles: [chat, edit]
+ capabilities: [thinking, tool_calling]
+ input_types: [text, image]
+ output_types: [text]
+ release_date: 2026-09-22
+ cost: {input: 0.1, output: 0.5}
+
- name: GPT-6 Astra (Global)
model: global.openai.gpt-6-astra
description: "OpenAI model for advanced analysis, software engineering, deep research, scientific work, and document creation"
@@ -422,7 +353,7 @@ bedrock:
- name: Claude Fable 5.1 (Global)
model: global.anthropic.claude-fable-5-1
- description: "Anthropic's most capable generally available model for ambitious, long-running coding, agentic, research, and enterprise workflows"
+ description: "Anthropic model for ambitious, long-running coding, agentic, research, and enterprise workflows"
roles: [chat, edit]
capabilities: [thinking, tool_calling]
input_types: [text, image, pdf]
@@ -550,36 +481,6 @@ bedrock:
release_date: 2026-05-28
cost: {input: 5, output: 25}
- - name: Claude Opus 4.7
- model: anthropic.claude-opus-4-7
- description: "Anthropic Opus model built for long-running, asynchronous agents"
- roles: [chat, edit]
- capabilities: [thinking, tool_calling]
- input_types: [text, image, pdf]
- output_types: [text]
- release_date: 2026-04-16
- cost: {input: 5, output: 25}
-
- - name: Claude Opus 4.7 (Global)
- model: global.anthropic.claude-opus-4-7
- description: "Anthropic Opus model built for long-running, asynchronous agents"
- roles: [chat, edit]
- capabilities: [thinking, tool_calling]
- input_types: [text, image, pdf]
- output_types: [text]
- release_date: 2026-04-16
- cost: {input: 5, output: 25}
-
- - name: MiniMax M2.5
- model: minimax.minimax-m2.5
- description: "Large language model designed for real-world productivity"
- roles: [chat, edit]
- capabilities: [thinking, tool_calling]
- input_types: [text]
- output_types: [text]
- release_date: 2026-03-18
- cost: {input: 0.3, output: 1.2}
-
- name: GLM-5
model: zai.glm-5
description: "Z.ai's flagship open-source foundation model engineered for complex systems design and long-horizon agent workflows"
@@ -611,6 +512,56 @@ bedrock:
cost: {input: 3, output: 15}
azure:
+ - name: Claude Opus 5.5
+ model: claude-opus-5-5
+ description: "Anthropic's strongest Opus model for coding, agents, and knowledge work"
+ roles: [chat, edit]
+ capabilities: [thinking, tool_calling]
+ input_types: [text, image, pdf]
+ output_types: [text]
+ release_date: 2026-09-22
+ cost: {input: 4, output: 20}
+
+ - name: GPT-6 Luna
+ model: gpt-6-luna
+ description: "OpenAI's most efficient model for focused, high-volume tasks"
+ roles: [chat, edit]
+ capabilities: [thinking, tool_calling]
+ input_types: [text, image, pdf]
+ output_types: [text]
+ release_date: 2026-09-22
+ cost: {input: 0.1, output: 0.5}
+
+ - name: GPT-6 Sol
+ model: gpt-6-sol
+ description: "OpenAI reasoning model for demanding tasks that balances capability, latency, and cost"
+ roles: [chat, edit]
+ capabilities: [thinking, tool_calling]
+ input_types: [text, image, pdf]
+ output_types: [text]
+ release_date: 2026-09-22
+ cost: {input: 2, output: 10}
+
+ - name: GPT Image 2.5 Flare
+ model: gpt-image-2.5-flare
+ description: "Small image generation and editing model optimized for speed and high-quality everyday use"
+ roles: [chat, edit]
+ capabilities: []
+ input_types: [text, image]
+ output_types: [image]
+ release_date: 2026-09-08
+ cost: {input: 5, output: 30}
+
+ - name: GPT Image 2.5 Sunburst
+ model: gpt-image-2.5-sunburst
+ description: "Image generation and editing model optimized for quality, precise editing, and subject preservation"
+ roles: [chat, edit]
+ capabilities: []
+ input_types: [text, image]
+ output_types: [image]
+ release_date: 2026-09-08
+ cost: {input: 5, output: 30}
+
- name: GPT-6 Astra
model: gpt-6-astra
description: "OpenAI model for advanced analysis, software engineering, deep research, scientific work, and document creation"
@@ -623,7 +574,7 @@ azure:
- name: Claude Fable 5.1
model: claude-fable-5-1
- description: "Anthropic's most capable generally available model for ambitious, long-running coding, agentic, research, and enterprise workflows"
+ description: "Anthropic model for ambitious, long-running coding, agentic, research, and enterprise workflows"
roles: [chat, edit]
capabilities: [thinking, tool_calling]
input_types: [text, image, pdf]
@@ -720,16 +671,6 @@ azure:
release_date: 2026-04-24
cost: {input: 5, output: 30}
- - name: Kimi K2.6
- model: kimi-k2.6
- description: "Moonshot AI multimodal model for long-horizon coding, coding-driven UI/UX generation, and multi-agent orchestration"
- roles: [chat, edit]
- capabilities: [thinking, tool_calling]
- input_types: [text, image]
- output_types: [text]
- release_date: 2026-04-22
- cost: {input: 0.95, output: 4}
-
- name: Grok 4.20 (Non-Reasoning)
model: grok-4-20-non-reasoning
description: "Reasoning model from xAI with industry-leading speed and agentic tool calling capabilities"
@@ -750,46 +691,6 @@ azure:
release_date: 2026-04-08
cost: {input: 2, output: 6}
- - name: GPT-5.4 Mini
- model: gpt-5.4-mini
- description: "GPT-5.4 mini brings the core capabilities of GPT-5.4 to a faster, more efficient model optimized for high-throughput workloads"
- roles: [chat, edit]
- capabilities: [thinking, tool_calling]
- input_types: [text, image, pdf]
- output_types: [text]
- release_date: 2026-03-17
- cost: {input: 0.75, output: 4.5}
-
- - name: GPT-5.4 Nano
- model: gpt-5.4-nano
- description: "Most lightweight and cost-efficient variant of the GPT-5.4 family, optimized for speed-critical and high-volume tasks"
- roles: [chat, edit]
- capabilities: [thinking, tool_calling]
- input_types: [text, image, pdf]
- output_types: [text]
- release_date: 2026-03-17
- cost: {input: 0.2, output: 1.25}
-
- - name: GPT-5.4
- model: gpt-5.4
- description: "OpenAI model that unified the Codex and GPT lines into a single system"
- roles: [chat, edit]
- capabilities: [thinking, tool_calling]
- input_types: [text, image, pdf]
- output_types: [text]
- release_date: 2026-03-05
- cost: {input: 2.5, output: 15}
-
- - name: GPT-5.3 Codex
- model: gpt-5.3-codex
- description: "Agentic coding model that combines GPT-5.2-Codex software engineering performance with GPT-5.2 reasoning and professional knowledge"
- roles: [chat, edit]
- capabilities: [thinking, tool_calling]
- input_types: [text, image]
- output_types: [text]
- release_date: 2026-02-24
- cost: {input: 1.75, output: 14}
-
- name: Claude Opus 4.8
model: claude-opus-4-8
description: "Modest but tangible improvement over Claude Opus 4.7"
@@ -903,24 +804,6 @@ ollama:
output_types: [text]
release_date: 2026-04-24
- - name: Kimi K2.7 Code
- model: kimi-k2.7-code
- description: "Coding-focused model in Moonshot AI's Kimi K2 family, built to complete end-to-end programming tasks reliably over long contexts"
- roles: [chat, edit]
- capabilities: [thinking, tool_calling]
- input_types: [text, image]
- output_types: [text]
- release_date: 2026-06-12
-
- - name: Kimi K2.6
- model: kimi-k2.6
- description: "Moonshot AI multimodal model for long-horizon coding, coding-driven UI/UX generation, and multi-agent orchestration"
- roles: [chat, edit]
- capabilities: [thinking, tool_calling]
- input_types: [text, image]
- output_types: [text]
- release_date: 2026-04-20
-
- name: Gemma 4 31B
model: gemma4:31b
description: "Google DeepMind's 30.7B dense multimodal model supporting text and image input with text output"
@@ -939,15 +822,6 @@ ollama:
output_types: [text]
release_date: 2026-06-13
- - name: GLM 5.1
- model: glm-5.1
- description: "GLM-5.1 delivers a major leap in coding capability, with particularly significant gains in handling long-horizon tasks"
- roles: [chat, edit]
- capabilities: [thinking, tool_calling]
- input_types: [text]
- output_types: [text]
- release_date: 2026-03-27
-
- name: MiniMax M3
model: minimax-m3
description: "Multimodal foundation model from MiniMax with a 1M-token context window, suited for long-horizon agentic work and coding"
@@ -994,6 +868,16 @@ ollama:
release_date: 2026-02-15
wandb:
+ - name: DeepSeek V4.1 Flash
+ model: deepseek-ai/DeepSeek-V4.1-Flash
+ description: "Sparse Mixture-of-Experts model from DeepSeek built on a Causal Encoder-Decoder architecture"
+ roles: [chat, edit]
+ capabilities: [thinking, tool_calling]
+ input_types: [text, image]
+ output_types: [text]
+ release_date: 2026-09-10
+ cost: {input: 0.2, output: 0.65}
+
- name: GLM 5.3 Flash
model: zai-org/GLM-5.3-Flash
description: "Native multimodal model from Z.ai optimized for efficient coding and long-horizon agent tasks"
@@ -1064,16 +948,6 @@ wandb:
release_date: 2026-06-12
cost: {input: 0.23, output: 0.96}
- - name: Granite 4.1 8B
- model: ibm-granite/granite-4.1-8b
- description: "Dense 8B decoder-only model from IBM's Granite 4.1 family with a 131K-token context window, designed for enterprise tasks"
- roles: [chat, edit]
- capabilities: [tool_calling]
- input_types: [text]
- output_types: [text]
- release_date: 2026-04-29
- cost: {input: 0.05, output: 0.1}
-
- name: DeepSeek V4-Flash
model: deepseek-ai/DeepSeek-V4-Flash
description: "DeepSeek V4-Flash is an MoE model with 1M context length great for coding, reasoning, and agentic workloads"
@@ -1104,16 +978,6 @@ wandb:
release_date: 2026-06-16
cost: {input: 0.76, output: 2.42}
- - name: Z.AI GLM-5.1
- model: zai-org/GLM-5.1
- description: "GLM-5.1 delivers a major leap in coding capability, with particularly significant gains in handling long-horizon tasks"
- roles: [chat, edit]
- capabilities: [thinking, tool_calling]
- input_types: [text]
- output_types: [text]
- release_date: 2026-04-07
- cost: {input: 1.4, output: 4.4}
-
- name: Nemotron 3 Ultra
model: nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B
description: "Open reasoning and orchestration model from NVIDIA with 55B active parameters out of 550B total (MoE)"
@@ -1134,87 +998,87 @@ wandb:
release_date: 2026-03-11
cost: {input: 0.2, output: 0.8}
- - name: MiniMax M2.5
- model: MiniMaxAI/MiniMax-M2.5
- description: "Large language model designed for real-world productivity"
+ - name: gpt-oss-120b
+ model: openai/gpt-oss-120b
+ description: "Open-weight, 117B-parameter Mixture-of-Experts (MoE) language model from OpenAI designed for high-reasoning, agentic, and general-purpose production use cases"
roles: [chat, edit]
capabilities: [thinking, tool_calling]
input_types: [text]
output_types: [text]
- release_date: 2026-02-12
- cost: {input: 0.3, output: 1.2}
+ release_date: 2025-08-05
+ cost: {input: 0.04, output: 0.14}
- - name: Kimi K2.7 Code
- model: moonshotai/Kimi-K2.7-Code
- description: "Kimi K2.7 Code is a 1T-parameter MoE model with 32B active parameters purpose-built for long-horizon agentic coding and software engineering"
+ - name: gpt-oss-20b
+ model: openai/gpt-oss-20b
+ description: "Open-weight 21B parameter model released by OpenAI under the Apache 2.0 license"
roles: [chat, edit]
capabilities: [thinking, tool_calling]
input_types: [text]
output_types: [text]
- release_date: 2026-06-12
- cost: {input: 0.71, output: 3.5}
+ release_date: 2025-08-05
+ cost: {input: 0.03, output: 0.13}
- - name: Moonshot AI Kimi K2.6
- model: moonshotai/Kimi-K2.6
- description: "Kimi K2.6 is a multimodal Mixture-of-Experts language model featuring 32 billion activated parameters and a total of 1 trillion parameters"
+ - name: Mellum2 12B A2.5B
+ model: JetBrains/Mellum2-12B-A2.5B-Instruct
+ description: "Mellum2-12B-A2.5B-Instruct is a fast MoE model with 131K context built for coding, tool use, and low-latency AI workflows"
roles: [chat, edit]
capabilities: [thinking, tool_calling]
- input_types: [text, image]
+ input_types: [text]
output_types: [text]
- release_date: 2026-04-20
- cost: {input: 0.95, output: 4}
+ release_date: 2026-06-01
+ cost: {input: 0.05, output: 0.1}
- - name: gpt-oss-120b
- model: openai/gpt-oss-120b
- description: "Open-weight, 117B-parameter Mixture-of-Experts (MoE) language model from OpenAI designed for high-reasoning, agentic, and general-purpose production use cases"
+opencode-go:
+ - name: LongCat 2.5 Preview Free
+ model: longcat-2.5-preview-free
+ description: "Multimodal reasoning model from Meituan for coding and agentic workflows, with a one-million-token context window"
roles: [chat, edit]
capabilities: [thinking, tool_calling]
- input_types: [text]
+ input_types: [text, image]
output_types: [text]
- release_date: 2025-08-05
- cost: {input: 0.04, output: 0.14}
+ release_date: 2026-09-25
+ cost: {input: 0, output: 0}
- - name: gpt-oss-20b
- model: openai/gpt-oss-20b
- description: "Open-weight 21B parameter model released by OpenAI under the Apache 2.0 license"
+ - name: Space Bunny Free
+ model: space-bunny-free
+ description: "Anonymous model with fast inference, strong coding capabilities, native multimodal input, and adjustable reasoning effort"
roles: [chat, edit]
capabilities: [thinking, tool_calling]
- input_types: [text]
+ input_types: [text, image]
output_types: [text]
- release_date: 2025-08-05
- cost: {input: 0.03, output: 0.13}
+ release_date: 2026-09-23
+ cost: {input: 0, output: 0}
- - name: Qwen3.6 35B A3B
- model: Qwen/Qwen3.6-35B-A3B
- description: "Qwen3.6-35B-A3B is an MoE multimodal model with 262K context optimized for agentic coding workflows"
+ - name: GPT-6 Luna
+ model: gpt-6-luna
+ description: "OpenAI's most efficient model for focused, high-volume tasks"
roles: [chat, edit]
capabilities: [thinking, tool_calling]
- input_types: [text, image]
+ input_types: [text, image, pdf]
output_types: [text]
- release_date: 2026-04-16
- cost: {input: 0.25, output: 1.25}
+ release_date: 2026-09-22
+ cost: {input: 0.1, output: 0.5}
- - name: Qwen3.6 27B
- model: Qwen/Qwen3.6-27B
- description: "Qwen3.6-27B is a 27B dense multimodal model with 262K context built for flagship-level agentic coding"
+ - name: MiMo-V2.6-Flash
+ model: mimo-v2.6-flash
+ description: "Open-source Mixture-of-Experts foundation model from Xiaomi with 309B total parameters and 15B activated per token"
roles: [chat, edit]
capabilities: [thinking, tool_calling]
input_types: [text, image]
output_types: [text]
- release_date: 2026-04-22
- cost: {input: 0.6, output: 3.6}
+ release_date: 2026-09-22
+ cost: {input: 0.14, output: 0.28}
- - name: Mellum2 12B A2.5B
- model: JetBrains/Mellum2-12B-A2.5B-Instruct
- description: "Mellum2-12B-A2.5B-Instruct is a fast MoE model with 131K context built for coding, tool use, and low-latency AI workflows"
+ - name: MiMo-V2.6-Pro
+ model: mimo-v2.6-pro
+ description: "Xiaomi's flagship foundation model with over one trillion parameters, designed for demanding coding and agentic tasks"
roles: [chat, edit]
capabilities: [thinking, tool_calling]
- input_types: [text]
+ input_types: [text, image]
output_types: [text]
- release_date: 2026-06-01
- cost: {input: 0.05, output: 0.1}
+ release_date: 2026-09-22
+ cost: {input: 0.435, output: 0.87}
-opencode-go:
- name: DeepSeek V4.1 Flash
model: deepseek-v4.1-flash
description: "Sparse Mixture-of-Experts model from DeepSeek built on a Causal Encoder-Decoder architecture"
@@ -1275,16 +1139,6 @@ opencode-go:
release_date: 2026-08-21
cost: {input: 0.22, output: 0.66}
- - name: Muse Spark 1.2 Contributor
- model: muse-spark-1.2-contributor
- description: "Discounted Meta coding model that permits prompts and completions to train future Meta models"
- roles: [chat, edit]
- capabilities: [thinking, tool_calling]
- input_types: [text, image, pdf]
- output_types: [text]
- release_date: 2026-08-05
- cost: {input: 0.1, output: 0.2}
-
- name: GLM-5.3
model: glm-5.3
description: "Flagship GLM model for long-horizon coding, agents, and complex project delivery"
@@ -1345,26 +1199,6 @@ opencode-go:
release_date: 2026-04-24
cost: {input: 1.74, output: 3.48}
- - name: Kimi K2.7 Code
- model: kimi-k2.7-code
- description: "Coding-focused model in Moonshot AI's Kimi K2 family, built to complete end-to-end programming tasks reliably over long contexts"
- roles: [chat, edit]
- capabilities: [thinking, tool_calling]
- input_types: [text, image]
- output_types: [text]
- release_date: 2026-06-12
- cost: {input: 0.95, output: 4}
-
- - name: Kimi K2.6
- model: kimi-k2.6
- description: "Moonshot AI multimodal model for long-horizon coding, coding-driven UI/UX generation, and multi-agent orchestration"
- roles: [chat, edit]
- capabilities: [thinking, tool_calling]
- input_types: [text, image]
- output_types: [text]
- release_date: 2026-04-21
- cost: {input: 0.95, output: 4}
-
- name: GLM 5.2
model: glm-5.2
description: "Large-scale reasoning model from Z.ai with a 1M-token context window, suited for long-horizon agent workflows and project-level software engineering"
@@ -1404,23 +1238,3 @@ opencode-go:
output_types: [text]
release_date: 2026-05-31
cost: {input: 0.1, output: 0.4}
-
- - name: MiMo V2.5
- model: mimo-v2.5
- description: "Native omnimodal model from Xiaomi delivering Pro-level agentic performance at roughly half the inference cost"
- roles: [chat, edit]
- capabilities: [thinking, tool_calling]
- input_types: [text, image]
- output_types: [text]
- release_date: 2026-04-22
- cost: {input: 0.4, output: 2}
-
- - name: MiMo V2.5 Pro
- model: mimo-v2.5-pro
- description: "Xiaomi's flagship model, delivering strong performance in general agentic capabilities, complex software engineering, and long-horizon tasks, with top rankings on benchmarks such as ClawEval, GDPVal,…"
- roles: [chat, edit]
- capabilities: [thinking, tool_calling]
- input_types: [text]
- output_types: [text]
- release_date: 2026-04-22
- cost: {input: 1, output: 3}