AI Release Timeline

Gemini 3.6 Flash launched by Google. Google's speed-optimized Gemini 3.6 Flash with frontier-level intelligence at lower cost than prior Flash tiers.

API endpoint `gemini-3.6-flash` available for Gemini 3.6 Flash.

Initial pricing set at $1.50/M input and $7.50/M output tokens.

Gemini 3.6 Flash ships with 1M token context window.

Gemini 3.6 Flash adds native vision and image understanding capabilities.

Gemini 3.6 Flash supports audio input and multimodal voice workflows.

Gemini 3.6 Flash supports function calling and structured tool use.

Gemini 3.6 Flash introduces extended reasoning and chain-of-thought capabilities.

Gemini 3.6 Flash achieves 94/100 multimodal benchmark score — top-tier performance.

Kimi K3 launched by Moonshot AI. Moonshot flagship open-frontier model with ~2.8T parameters, native vision, and a 1M-token context window.

API endpoint `kimi-k3` available for Kimi K3.

Initial pricing set at $3.00/M input and $15.00/M output tokens.

Kimi K3 ships with 1M token context window.

Kimi K3 adds native vision and image understanding capabilities.

Kimi K3 supports function calling and structured tool use.

Kimi K3 introduces extended reasoning and chain-of-thought capabilities.

Kimi K3 achieves 94/100 reasoning benchmark score — top-tier performance.

Claude Sonnet 4.5 launched by Anthropic. Anthropic's balanced Sonnet 4.5 for production coding and agent workflows.

API endpoint `claude-sonnet-4-5` available for Claude Sonnet 4.5.

Initial pricing set at $3.00/M input and $15.00/M output tokens.

Claude Sonnet 4.5 adds native vision and image understanding capabilities.

Claude Sonnet 4.5 supports function calling and structured tool use.

Claude Sonnet 4.5 introduces extended reasoning and chain-of-thought capabilities.

Claude Sonnet 4.5 achieves 95/100 coding benchmark score — top-tier performance.

GPT-5 launched by OpenAI. OpenAI flagship GPT-5 generation for general and agentic workloads.

API endpoint `gpt-5` available for GPT-5.

Initial pricing set at $5.00/M input and $20.00/M output tokens.

GPT-5 adds native vision and image understanding capabilities.

GPT-5 supports audio input and multimodal voice workflows.

GPT-5 supports function calling and structured tool use.

GPT-5 introduces extended reasoning and chain-of-thought capabilities.

GPT-5 achieves 96/100 reasoning benchmark score — top-tier performance.

GPT-5 Mini launched by OpenAI. Cost-efficient GPT-5 family model for high-volume production.

API endpoint `gpt-5-mini` available for GPT-5 Mini.

Initial pricing set at $0.40/M input and $1.60/M output tokens.

GPT-5 Mini adds native vision and image understanding capabilities.

GPT-5 Mini supports function calling and structured tool use.

GPT-5 Mini introduces extended reasoning and chain-of-thought capabilities.

DeepSeek V4 launched by DeepSeek. Next-gen DeepSeek general model emphasizing coding and cost efficiency.

API endpoint `deepseek-v4` available for DeepSeek V4.

Initial pricing set at $0.30/M input and $1.20/M output tokens.

DeepSeek V4 supports function calling and structured tool use.

DeepSeek V4 introduces extended reasoning and chain-of-thought capabilities.

DeepSeek V4 achieves 92/100 coding benchmark score — top-tier performance.

DeepSeek V4 available for fine-tuning and custom deployment via DeepSeek.

Claude Opus 4.1 launched by Anthropic. Anthropic's latest Opus-class model for deep reasoning, coding, and long-form analysis.

API endpoint `claude-opus-4-1` available for Claude Opus 4.1.

Initial pricing set at $15.00/M input and $75.00/M output tokens.

Claude Opus 4.1 adds native vision and image understanding capabilities.

Claude Opus 4.1 supports function calling and structured tool use.

Claude Opus 4.1 introduces extended reasoning and chain-of-thought capabilities.

Claude Opus 4.1 achieves 97/100 reasoning benchmark score — top-tier performance.

Gemini 3 Pro launched by Google. Google Gemini 3 Pro frontier tier for complex multimodal reasoning and long-context workloads.

API endpoint `gemini-3-pro` available for Gemini 3 Pro.

Initial pricing set at $3.60/M input and $14.00/M output tokens.

Gemini 3 Pro ships with 1M token context window.

Gemini 3 Pro adds native vision and image understanding capabilities.

Gemini 3 Pro supports audio input and multimodal voice workflows.

Gemini 3 Pro supports function calling and structured tool use.

Gemini 3 Pro introduces extended reasoning and chain-of-thought capabilities.

Gemini 3 Pro achieves 96/100 multimodal benchmark score — top-tier performance.

Grok 4 launched by xAI. xAI Grok 4 with real-time retrieval emphasis and strong general capability.

API endpoint `grok-4` available for Grok 4.

Initial pricing set at $3.00/M input and $15.00/M output tokens.

Grok 4 adds native vision and image understanding capabilities.

Grok 4 supports function calling and structured tool use.

Grok 4 introduces extended reasoning and chain-of-thought capabilities.

Kimi K2 launched by Moonshot AI. Prior Moonshot coding-strong MoE model, strong agent/tool use at competitive pricing.

API endpoint `kimi-k2` available for Kimi K2.

Initial pricing set at $0.60/M input and $2.50/M output tokens.

Kimi K2 supports function calling and structured tool use.

Kimi K2 introduces extended reasoning and chain-of-thought capabilities.

Claude 4 Sonnet launched by Anthropic. Latest Sonnet with improved reasoning and tool use.

API endpoint `claude-sonnet-4-20250514` available for Claude 4 Sonnet.

Initial pricing set at $3.00/M input and $15.00/M output tokens.

Claude 4 Sonnet adds native vision and image understanding capabilities.

Claude 4 Sonnet supports function calling and structured tool use.

Claude 4 Sonnet introduces extended reasoning and chain-of-thought capabilities.

Claude 4 Sonnet achieves 96/100 coding benchmark score — top-tier performance.

Claude 4 Opus launched by Anthropic. Anthropic flagship for the most demanding tasks.

API endpoint `claude-opus-4-20250514` available for Claude 4 Opus.

Initial pricing set at $15.00/M input and $75.00/M output tokens.

Claude 4 Opus adds native vision and image understanding capabilities.

Claude 4 Opus supports function calling and structured tool use.

Claude 4 Opus introduces extended reasoning and chain-of-thought capabilities.

Claude 4 Opus achieves 97/100 coding benchmark score — top-tier performance.

Claude 4 Haiku launched by Anthropic. Next-gen Haiku with Claude 4 intelligence at low cost.

API endpoint `claude-haiku-4-20250514` available for Claude 4 Haiku.

Initial pricing set at $0.80/M input and $4.00/M output tokens.

Claude 4 Haiku adds native vision and image understanding capabilities.

Claude 4 Haiku supports function calling and structured tool use.

Gemini 2.5 Flash launched by Google. Flash model with thinking capabilities and top multimodal scores.

API endpoint `gemini-2.5-flash-preview-05-20` available for Gemini 2.5 Flash.

Initial pricing set at $0.15/M input and $0.60/M output tokens.

Gemini 2.5 Flash ships with 1M token context window.

Gemini 2.5 Flash adds native vision and image understanding capabilities.

Gemini 2.5 Flash supports audio input and multimodal voice workflows.

Gemini 2.5 Flash supports function calling and structured tool use.

Gemini 2.5 Flash introduces extended reasoning and chain-of-thought capabilities.

Gemini 2.5 Flash achieves 94/100 multimodal benchmark score — top-tier performance.

OpenAI o4 Mini launched by OpenAI. Latest compact reasoning model with vision support.

API endpoint `o4-mini` available for OpenAI o4 Mini.

Initial pricing set at $1.10/M input and $4.40/M output tokens.

OpenAI o4 Mini adds native vision and image understanding capabilities.

OpenAI o4 Mini supports function calling and structured tool use.

OpenAI o4 Mini introduces extended reasoning and chain-of-thought capabilities.

OpenAI o4 Mini achieves 93/100 reasoning benchmark score — top-tier performance.

OpenAI o3 launched by OpenAI. Full o3 reasoning model with vision and tool use.

API endpoint `o3` available for OpenAI o3.

Initial pricing set at $10.00/M input and $40.00/M output tokens.

OpenAI o3 supports function calling and structured tool use.

OpenAI o3 introduces extended reasoning and chain-of-thought capabilities.

OpenAI o3 achieves 97/100 reasoning benchmark score — top-tier performance.

GPT-4.1 launched by OpenAI. Next-gen GPT with 1M context and improved instruction following.

API endpoint `gpt-4.1` available for GPT-4.1.

Initial pricing set at $2.00/M input and $8.00/M output tokens.

GPT-4.1 ships with 1M token context window.

GPT-4.1 adds native vision and image understanding capabilities.

GPT-4.1 supports audio input and multimodal voice workflows.

GPT-4.1 supports function calling and structured tool use.

GPT-4.1 achieves 94/100 coding benchmark score — top-tier performance.

GPT-4.1 Mini launched by OpenAI. Mini variant of GPT-4.1 for scalable deployments.

API endpoint `gpt-4.1-mini` available for GPT-4.1 Mini.

Initial pricing set at $0.40/M input and $1.60/M output tokens.

GPT-4.1 Mini ships with 1M token context window.

GPT-4.1 Mini adds native vision and image understanding capabilities.

GPT-4.1 Mini supports audio input and multimodal voice workflows.

GPT-4.1 Mini supports function calling and structured tool use.

GPT-4.1 Nano launched by OpenAI. Ultra-low-cost GPT-4.1 tier for classification and extraction.

API endpoint `gpt-4.1-nano` available for GPT-4.1 Nano.

Initial pricing set at $0.10/M input and $0.40/M output tokens.

GPT-4.1 Nano ships with 1M token context window.

GPT-4.1 Nano adds native vision and image understanding capabilities.

GPT-4.1 Nano supports function calling and structured tool use.

Gemini 2.5 Pro launched by Google. Google flagship with 1M context and deep reasoning.

API endpoint `gemini-2.5-pro-preview-06-05` available for Gemini 2.5 Pro.

Initial pricing set at $1.25/M input and $10.00/M output tokens.

Gemini 2.5 Pro ships with 1M token context window.

Gemini 2.5 Pro adds native vision and image understanding capabilities.

Gemini 2.5 Pro supports audio input and multimodal voice workflows.

Gemini 2.5 Pro supports function calling and structured tool use.

Gemini 2.5 Pro introduces extended reasoning and chain-of-thought capabilities.

Gemini 2.5 Pro achieves 97/100 multimodal benchmark score — top-tier performance.

DeepSeek V3 0324 launched by DeepSeek. Updated V3 with improved instruction following and 128K context.

API endpoint `deepseek-chat` available for DeepSeek V3 0324.

Initial pricing set at $0.27/M input and $1.10/M output tokens.

DeepSeek V3 0324 supports function calling and structured tool use.

DeepSeek V3 0324 achieves 92/100 coding benchmark score — top-tier performance.

DeepSeek V3 0324 available for fine-tuning and custom deployment via DeepSeek.

Command A launched by Cohere. Cohere flagship agentic model for tool use and search.

API endpoint `command-a-03-2025` available for Command A.

Initial pricing set at $2.50/M input and $10.00/M output tokens.

Command A adds native vision and image understanding capabilities.

Command A supports function calling and structured tool use.

Llama 4 Maverick launched by Meta. Open multimodal Llama with 1M context window.

API endpoint `llama-4-maverick` available for Llama 4 Maverick.

Initial pricing set at $0.20/M input and $0.60/M output tokens.

Llama 4 Maverick ships with 1M token context window.

Llama 4 Maverick adds native vision and image understanding capabilities.

Llama 4 Maverick supports function calling and structured tool use.

Llama 4 Maverick available for fine-tuning and custom deployment via Meta.

Llama 4 Scout launched by Meta. Llama 4 with industry-leading 10M context for documents.

API endpoint `llama-4-scout` available for Llama 4 Scout.

Initial pricing set at $0.15/M input and $0.50/M output tokens.

Llama 4 Scout ships with 10M token context window.

Llama 4 Scout adds native vision and image understanding capabilities.

Llama 4 Scout supports function calling and structured tool use.

Llama 4 Scout available for fine-tuning and custom deployment via Meta.

Claude 3.7 Sonnet launched by Anthropic. Sonnet with extended thinking for harder problems.

API endpoint `claude-3-7-sonnet-20250219` available for Claude 3.7 Sonnet.

Initial pricing set at $3.00/M input and $15.00/M output tokens.

Claude 3.7 Sonnet adds native vision and image understanding capabilities.

Claude 3.7 Sonnet supports function calling and structured tool use.

Claude 3.7 Sonnet introduces extended reasoning and chain-of-thought capabilities.

Claude 3.7 Sonnet achieves 95/100 coding benchmark score — top-tier performance.

Grok 3 launched by xAI. xAI flagship with strong reasoning and tool use.

API endpoint `grok-3` available for Grok 3.

Initial pricing set at $3.00/M input and $15.00/M output tokens.

Grok 3 adds native vision and image understanding capabilities.

Grok 3 supports function calling and structured tool use.

Grok 3 introduces extended reasoning and chain-of-thought capabilities.

Grok 3 achieves 92/100 reasoning benchmark score — top-tier performance.

Grok 3 Mini launched by xAI. Efficient Grok 3 variant for high-volume use.

API endpoint `grok-3-mini` available for Grok 3 Mini.

Initial pricing set at $0.30/M input and $0.50/M output tokens.

Grok 3 Mini adds native vision and image understanding capabilities.

Grok 3 Mini supports function calling and structured tool use.

Grok 3 Mini introduces extended reasoning and chain-of-thought capabilities.

Gemini 2.0 Flash Lite launched by Google. Ultra-cheap Gemini for simple tasks at scale.

API endpoint `gemini-2.0-flash-lite` available for Gemini 2.0 Flash Lite.

Initial pricing set at $0.075/M input and $0.30/M output tokens.

Gemini 2.0 Flash Lite ships with 1M token context window.

Gemini 2.0 Flash Lite adds native vision and image understanding capabilities.

Gemini 2.0 Flash Lite supports function calling and structured tool use.

OpenAI o3 Mini launched by OpenAI. Efficient reasoning model in the o3 family.

API endpoint `o3-mini` available for OpenAI o3 Mini.

Initial pricing set at $1.10/M input and $4.40/M output tokens.

OpenAI o3 Mini supports function calling and structured tool use.

OpenAI o3 Mini introduces extended reasoning and chain-of-thought capabilities.

Mistral Small launched by Mistral. Compact Mistral for cost-sensitive applications.

API endpoint `mistral-small-2501` available for Mistral Small.

Initial pricing set at $0.10/M input and $0.30/M output tokens.

Mistral Small supports function calling and structured tool use.

Mistral Small available for fine-tuning and custom deployment via Mistral.

DeepSeek R1 launched by DeepSeek. Open reasoning model with o1-class performance at fraction of cost.

API endpoint `deepseek-reasoner` available for DeepSeek R1.

Initial pricing set at $0.55/M input and $2.19/M output tokens.

DeepSeek R1 supports function calling and structured tool use.

DeepSeek R1 introduces extended reasoning and chain-of-thought capabilities.

DeepSeek R1 achieves 95/100 reasoning benchmark score — top-tier performance.

DeepSeek R1 available for fine-tuning and custom deployment via DeepSeek.

Codestral launched by Mistral. Code-specialized Mistral model for IDE and agent use.

API endpoint `codestral-2501` available for Codestral.

Initial pricing set at $0.30/M input and $0.90/M output tokens.

Codestral supports function calling and structured tool use.

Codestral achieves 94/100 coding benchmark score — top-tier performance.

Codestral available for fine-tuning and custom deployment via Mistral.

DeepSeek V3 launched by DeepSeek. Highly efficient MoE model rivaling frontier closed models.

API endpoint `deepseek-chat` available for DeepSeek V3.

Initial pricing set at $0.27/M input and $1.10/M output tokens.

DeepSeek V3 supports function calling and structured tool use.

DeepSeek V3 available for fine-tuning and custom deployment via DeepSeek.

Gemini 2.0 Flash launched by Google. Fast multimodal Gemini with native audio and vision.

API endpoint `gemini-2.0-flash` available for Gemini 2.0 Flash.

Initial pricing set at $0.10/M input and $0.40/M output tokens.

Gemini 2.0 Flash ships with 1M token context window.

Gemini 2.0 Flash adds native vision and image understanding capabilities.

Gemini 2.0 Flash supports audio input and multimodal voice workflows.

Gemini 2.0 Flash supports function calling and structured tool use.

Gemini 2.0 Flash achieves 92/100 multimodal benchmark score — top-tier performance.

Llama 3.3 70B launched by Meta. Open-weight 70B model competitive with larger closed models.

API endpoint `llama-3.3-70b-instruct` available for Llama 3.3 70B.

Initial pricing set at $0.23/M input and $0.40/M output tokens.

Llama 3.3 70B supports function calling and structured tool use.

Llama 3.3 70B available for fine-tuning and custom deployment via Meta.

OpenAI o1 launched by OpenAI. Reasoning model using internal chain-of-thought.

API endpoint `o1` available for OpenAI o1.

Initial pricing set at $15.00/M input and $60.00/M output tokens.

OpenAI o1 supports function calling and structured tool use.

OpenAI o1 introduces extended reasoning and chain-of-thought capabilities.

OpenAI o1 achieves 96/100 reasoning benchmark score — top-tier performance.

Mistral Large launched by Mistral. Mistral flagship for enterprise European deployments.

API endpoint `mistral-large-2411` available for Mistral Large.

Initial pricing set at $2.00/M input and $6.00/M output tokens.

Mistral Large adds native vision and image understanding capabilities.

Mistral Large supports function calling and structured tool use.

Mistral Large available for fine-tuning and custom deployment via Mistral.

Pixtral Large launched by Mistral. Mistral multimodal model with strong vision understanding.

API endpoint `pixtral-large-2411` available for Pixtral Large.

Initial pricing set at $2.00/M input and $6.00/M output tokens.

Pixtral Large adds native vision and image understanding capabilities.

Pixtral Large supports function calling and structured tool use.

Pixtral Large achieves 93/100 multimodal benchmark score — top-tier performance.

Pixtral Large available for fine-tuning and custom deployment via Mistral.

Claude 3.5 Haiku launched by Anthropic. Fast and affordable Claude for high-throughput workloads.

API endpoint `claude-3-5-haiku-20241022` available for Claude 3.5 Haiku.

Initial pricing set at $0.80/M input and $4.00/M output tokens.

Claude 3.5 Haiku adds native vision and image understanding capabilities.

Claude 3.5 Haiku supports function calling and structured tool use.

Ministral 8B launched by Mistral. Tiny on-device capable Mistral for edge deployments.

API endpoint `ministral-8b-2410` available for Ministral 8B.

Initial pricing set at $0.10/M input and $0.10/M output tokens.

Ministral 8B supports function calling and structured tool use.

Ministral 8B available for fine-tuning and custom deployment via Mistral.

Grok 2 launched by xAI. xAI general model with real-time X data integration.

API endpoint `grok-2` available for Grok 2.

Initial pricing set at $2.00/M input and $10.00/M output tokens.

Grok 2 adds native vision and image understanding capabilities.

Grok 2 supports function calling and structured tool use.

Llama 3.1 405B launched by Meta. Largest open-weight Llama for maximum capability.

API endpoint `llama-3.1-405b-instruct` available for Llama 3.1 405B.

Initial pricing set at $0.35/M input and $0.70/M output tokens.

Llama 3.1 405B supports function calling and structured tool use.

Llama 3.1 405B available for fine-tuning and custom deployment via Meta.

GPT-4o Mini launched by OpenAI. Cost-efficient small model for high-volume applications.

API endpoint `gpt-4o-mini` available for GPT-4o Mini.

Initial pricing set at $0.15/M input and $0.60/M output tokens.

GPT-4o Mini adds native vision and image understanding capabilities.

GPT-4o Mini supports function calling and structured tool use.

Claude 3.5 Sonnet launched by Anthropic. Balanced Claude model excelling at coding and analysis.

API endpoint `claude-3-5-sonnet-20241022` available for Claude 3.5 Sonnet.

Initial pricing set at $3.00/M input and $15.00/M output tokens.

Claude 3.5 Sonnet adds native vision and image understanding capabilities.

Claude 3.5 Sonnet supports function calling and structured tool use.

Claude 3.5 Sonnet achieves 94/100 writing benchmark score — top-tier performance.

DeepSeek Coder V2 launched by DeepSeek. Code-specialized DeepSeek for programming tasks.

API endpoint `deepseek-coder` available for DeepSeek Coder V2.

Initial pricing set at $0.14/M input and $0.28/M output tokens.

DeepSeek Coder V2 supports function calling and structured tool use.

DeepSeek Coder V2 achieves 93/100 coding benchmark score — top-tier performance.

DeepSeek Coder V2 available for fine-tuning and custom deployment via DeepSeek.

GPT-4o launched by OpenAI. Flagship multimodal model balancing speed, intelligence, and cost.

API endpoint `gpt-4o` available for GPT-4o.

Initial pricing set at $2.50/M input and $10.00/M output tokens.

GPT-4o adds native vision and image understanding capabilities.

GPT-4o supports audio input and multimodal voice workflows.

GPT-4o supports function calling and structured tool use.

GPT-4o achieves 95/100 multimodal benchmark score — top-tier performance.

Command R+ launched by Cohere. Cohere model optimized for RAG and enterprise search.

API endpoint `command-r-plus` available for Command R+.

Initial pricing set at $2.50/M input and $10.00/M output tokens.

Command R+ supports function calling and structured tool use.

Command R launched by Cohere. Affordable Cohere model for retrieval-augmented workflows.

API endpoint `command-r` available for Command R.

Initial pricing set at $0.15/M input and $0.60/M output tokens.

Command R supports function calling and structured tool use.

Claude 3 Opus launched by Anthropic. Previous-gen Opus still strong for writing-heavy workloads.

API endpoint `claude-3-opus-20240229` available for Claude 3 Opus.

Initial pricing set at $15.00/M input and $75.00/M output tokens.

Claude 3 Opus adds native vision and image understanding capabilities.

Claude 3 Opus supports function calling and structured tool use.

Claude 3 Opus achieves 93/100 writing benchmark score — top-tier performance.

Gemini 1.5 Pro launched by Google. Proven Gemini with industry-leading 2M context window.

API endpoint `gemini-1.5-pro` available for Gemini 1.5 Pro.

Initial pricing set at $1.25/M input and $5.00/M output tokens.

Gemini 1.5 Pro ships with 2.1M token context window.

Gemini 1.5 Pro adds native vision and image understanding capabilities.

Gemini 1.5 Pro supports function calling and structured tool use.