LLM Models

Anthropic's fastest model with near-frontier intelligence (October 2025). Supports extended thinking and delivers strong coding and agentic performance at the lowest Claude pricing tier.

September 2025 upgrade to the Sonnet line with stronger coding and agentic performance and a 1M-token context window (beta). The balanced choice for most production workloads.

DISCONTINUED. An August 2025 incremental upgrade to Claude Opus 4 with improved agentic and coding performance. Anthropic has deprecated the model and retires it on August 5, 2026; Claude Opus 5 is the documented migration target. Superseded on price and capability by the Opus 4.5 generation onward.

The balanced workhorse of Anthropic's first Claude 4 generation (May 2025), pairing strong intelligence with fast, cost-effective inference. Deprecated and scheduled for retirement on June 15, 2026 in favor of Claude Sonnet 4.6.

The flagship of Anthropic's first Claude 4 generation (May 2025), built for complex reasoning and long-horizon agentic coding. Deprecated and scheduled for retirement on June 15, 2026 in favor of Claude Opus 4.7.

Anthropic's first hybrid reasoning model, combining quick responses with optional extended thinking in a single model. Strong gains in coding and agentic tasks, and introduced configurable thinking budgets. Superseded by the Claude 4 family.

Anthropic's mid-2024 flagship Sonnet model and the first to introduce computer use (beta). Strong coding and agentic performance at a fraction of Opus pricing. The October 2024 revision improved tool use and agentic coding. Superseded by the Claude 4 family.

Anthropic's fastest model that matches Claude 3 Opus performance at similar speed to previous-gen Haiku. Released October 2024 with July 2024 training cutoff. Achieves 40.6% on SWE-bench Verified, outperforming agents using Claude 3.5 Sonnet and GPT-4o. Features 200K context window and 8K max output. Surpasses Claude 3 Opus on intelligence benchmarks while maintaining Haiku-class speed. Pricing: $0.80 per 1M input, $4 per 1M output (revised Dec 2024). Supports prompt caching (90% savings) and Message Batches API (50% savings).

Anthropic's fastest and most affordable model in its intelligence class. Processes 21K tokens (~30 pages) per second for prompts under 32K tokens, making it 3x faster than peers. Features 200K context window, vision capabilities, and surprisingly strong performance - during testing, surpassed the previous flagship Claude 3 Opus on many benchmarks. Exceptional value proposition: can process 400 Supreme Court cases or 2,500 images for just $1. State-of-the-art speed-to-intelligence ratio.

Anthropic's most capable model in the Claude 3 family, setting new standards for AI performance. Outperforms peers on most common evaluation benchmarks including graduate-level reasoning (GPQA), undergraduate knowledge (MMLU), and basic math (GSM8K). Achieved near-perfect 99% accuracy on 'Needle in a Haystack' recall evaluation. Features sophisticated vision capabilities for analyzing photos, charts, and technical diagrams. 200K context window expandable to 1M for specific use cases. Note: Deprecated as of January 2026, scheduled for retirement January 5, 2026 - migrate to Claude Opus 4.1.

OpenAI's cost-efficient reasoning model optimized for STEM tasks. Uses same high-compute reinforcement learning pipeline as o1-preview but with smaller, faster model. Achieves 70% on AIME math competition (competitive with o1's 74.4%), placing it in top 500 US high school students. Reaches 1650 Elo on Codeforces (86th percentile of programmers), nearly matching o1's 1673 Elo. Generates answers 3-5x faster than o1-preview while being 80% cheaper. Excels at STEM reasoning but performs worse on non-STEM factual knowledge. Recommended upgrade to o3-mini available.

OpenAI's most cost-efficient small model delivering intelligent performance at breakthrough pricing. Despite compact size, surpasses GPT-3.5 Turbo across academic benchmarks including 82% on MMLU (textual intelligence), 87% on MGSM (math reasoning), and 87.2% on HumanEval (coding). Features 128K context window and supports both text and vision inputs. Priced at 15 cents per 1M input tokens (60% cheaper than GPT-3.5 Turbo), making it ideal for high-volume applications requiring reliable intelligence without flagship costs.