Anthropic's current smallest and fastest Claude model, released October 2026 for high-volume, latency-sensitive work such as classification, extraction, routing and subagent tasks. The first Haiku with an effort control, plus a 1M-token context.
Mid-tier GPT-6 model released 2026-09-29 that OpenAI says nears GPT-6 Astra on coding, computer use and professional work at a fifth of Astra's token price; 1.05M context, low-to-max effort, optional Ultrafast mode.
Anthropic's current Sonnet model, released September 2026 as a faster, lower-cost complement to Claude Opus 5.5. Balances speed and intelligence with a 1M-token context, adaptive thinking and a between_tools setting that skips up-front thinking.
Anthropic's current Opus model and its recommended starting point for most workloads as of October 2026, first of the Claude 5.5 family. Built for long-running agentic coding and knowledge work, with always-on adaptive thinking and a 1M-token context.
Most efficient GPT-6 model (September 2026) for focused, high-volume work, priced at $0.10 input and $0.50 output per million tokens, with the full 1.05M context and effort from none to max. OpenAI's suggested replacement for retiring GPT-4o and GPT-4.1 snapshots.
First Sol model of the GPT-6 generation (September 2026), built for complex coding and agentic workflows with a 1.05M context and none-to-max reasoning. Followed a week later by GPT-6.1 Sol, which replaced it in OpenAI's main lineup.
xAI's flagship model as of October 2026, released September 21, 2026 for coding, agents and knowledge work. Offers a 500k-token context, no text output limit and reasoning effort up to xhigh.
Now served by DeepSeek-V4.1-Flash (API name deepseek-flash), DeepSeek's newest model and cheaper tier since September 2026: an open-weight 552B MoE with a 1M context. It replaced the retired V4-Flash.
DeepSeek-V4.1-Flash (September 2026), the first model on DeepSeek's V4.1 architecture: it uses a 552B backbone activating 8B parameters in prefill and 16B in decode, and supports 1M tokens under MIT.
OpenAI's flagship as of October 2026, released September 2026 as the top GPT-6 model for demanding reasoning, coding, computer use and research. 1.05M-token context, 128K output, reasoning effort from low to max.
Google's current Flash model, GA since September 2026, aimed at long-horizon software engineering and autonomous agents. Reads text, image, video, audio and PDF in a 1M-token context, with 64K output and low, medium or high thinking.
Upgraded September 2, 2026 snapshot of Qwen3.8-Max (also named qwen3.8-max-2026-09-02) with better engineering-scale coding and multi-tool agent orchestration, keeping the 1M context. The newest Max build.
Anthropic's most capable generally available model as of October 2026, succeeding Claude Fable 5. A Mythos-class tier above Opus for demanding reasoning and long-horizon agentic work, with always-on adaptive thinking and a 1M-token context.
The VL extension of Ling-3.0-flash (September 2026), served here for text input only; 124B MoE with about 5.5B active and a context of up to 1M tokens, thinking on by default.
Current Qwen Flash tier (August 2026): a model with a 1M-token context for coding assistance and agent workflows, compatible with OpenAI and Anthropic API formats; Alibaba's low-cost recommendation.
GLM-5.3-Flash (August 2026), a newly trained Z.ai GLM-5 model: a 320B MoE with 18B active, hybrid sparse and linear attention and MIT weights, which Z.ai says beats GLM-5.2 at about a tenth of the price.
Mid-size Granite 4.2 model (dense 8B, August 2026) from IBM, with built-in chain-of-thought that can be switched to low-effort or off, tool calling for agent workflows and a 128K native context; Apache 2.0, aimed at cost-conscious enterprise use.
Z.ai's GLM-5.3 (August 2026), GLM-5.2's base with heavier post-training for coding and agentic work and a 1M-token context, served by Alibaba Cloud, Mistral and Together.
Qwen3.8 27B dense model (August 2026, open weights), improving on 3.6-27B in coding and office tasks; 1M context on Model Studio. The newest dense Qwen model.
The August 2026 general-availability release of DeepSeek-V4-Pro, a 1.6T-parameter MoE with 49B active and a 1M-token context, served by DeepInfra and Together with low, high and max reasoning effort.
