Recursive: Flint (free) Flint (free) 16.4B tokens Flint is an open-weight mixture-of-experts model from Recursive, with about 2.5B active parameters out of 40B total, built by Recursive's autoresearch system Eureka Machine for coding agents. It is suited for investigating codebases, fixing bugs, implementing features, and writing tests. by recursive Oct 21, 2026 262K context $0 /M input tokens $0 /M output tokens Flint is an open-weight mixture-of-experts model from Recursive, with about 2.5B active parameters out of 40B total, built by Recursive's autoresearch system Eureka Machine for coding agents. It is suited for investigating codebases, fixing bugs, implementing features, and writing tests. Baseten: Mammoth 1 (free) Mammoth 1 (free) 71M tokens Mammoth 1 is a 27B decision model from Base Labs, Baseten's research team, built on Qwen3.5-27B. Given a state (text or JSON) and typed questions (yes/no, multiple choice, or ordinal score), it returns a calibrated probability for every option in a single forward pass, with no text generation. It is TypeSafe System One compatible and suits classification, routing, and scoring where you need a scored answer instead of prose. Its smaller sibling is Elephant 1 . by baseten Oct 21, 2026 $0 /M input tokens $0 /M output tokens Mammoth 1 is a 27B decision model from Base Labs, Baseten's research team, built on Qwen3.5-27B. Given a state (text or JSON) and typed questions (yes/no, multiple choice, or ordinal score), it returns a calibrated probability for every option in a single forward pass, with no text generation. It is TypeSafe System One compatible and suits classification, routing, and scoring where you need a scored answer instead of prose. Its smaller sibling is Elephant 1 . Baseten: Elephant 1 (free) Elephant 1 (free) 45.4M tokens Elephant 1 is a 9B decision model from Base Labs, Baseten's research team, built on Qwen3.5-9B. Given a state (text or JSON) and typed questions (yes/no, multiple choice, or ordinal score), it returns a calibrated probability for every option in a single forward pass, with no text generation. It is TypeSafe System One compatible and suits routing, classification, and scoring where you need a scored answer instead of prose. Its larger sibling is Mammoth 1 . by baseten Oct 21, 2026 $0 /M input tokens $0 /M output tokens Elephant 1 is a 9B decision model from Base Labs, Baseten's research team, built on Qwen3.5-9B. Given a state (text or JSON) and typed questions (yes/no, multiple choice, or ordinal score), it returns a calibrated probability for every option in a single forward pass, with no text generation. It is TypeSafe System One compatible and suits routing, classification, and scoring where you need a scored answer instead of prose. Its larger sibling is Mammoth 1 . Nace.AI: Drex Max Drex Max 284M tokens Drex Max is the largest decision model from Nace.AI, the successor to Drex v2.0 . Given a state (text or JSON) and typed questions (yes/no, multiple choice, or ordinal score), it returns a calibrated probability for every option in a single forward pass, with no text generation. It is TypeSafe System One compatible and suits routing, classification, and policy checks that need a scored answer instead of prose. by nace-ai Oct 21, 2026 $0.06 /M input tokens $0 /M output tokens Drex Max is the largest decision model from Nace.AI, the successor to Drex v2.0 . Given a state (text or JSON) and typed questions (yes/no, multiple choice, or ordinal score), it returns a calibrated probability for every option in a single forward pass, with no text generation. It is TypeSafe System One compatible and suits routing, classification, and policy checks that need a scored answer instead of prose. NVIDIA: Nemotron 3.5 Embed 1B (free) Nemotron 3.5 Embed 1B (free) 25.7M tokens Nemotron 3.5 Embed 1B is a multimodal embedding model from NVIDIA for multilingual retrieval over text and document page images (tables, charts, infographics). Queries are text and documents can be text, an image, or both. It returns 2048-dim vectors that can be truncated, and shares the text embedding space of Nemotron 3 Embed 1B. by nvidia Oct 21, 2026 $0 /M input tokens $0 /M output tokens Nemotron 3.5 Embed 1B is a multimodal embedding model from NVIDIA for multilingual retrieval over text and document page images (tables, charts, infographics). Queries are text and documents can be text, an image, or both. It returns 2048-dim vectors that can be truncated, and shares the text embedding space of Nemotron 3 Embed 1B. NVIDIA: Nemotron 3.5 Embed 8B (free) Nemotron 3.5 Embed 8B (free) 70.7M tokens Nemotron 3.5 Embed 8B is a multimodal embedding model from NVIDIA for multilingual retrieval over text and document page images (tables, charts, infographics). Queries are text and documents can be text, an image, or both. It returns 4096-dim vectors that can be truncated, and shares the text embedding space of Nemotron 3 Embed 8B. by nvidia Oct 21, 2026 $0 /M input tokens $0 /M output tokens Nemotron 3.5 Embed 8B is a multimodal embedding model from NVIDIA for multilingual retrieval over text and document page images (tables, charts, infographics). Queries are text and documents can be text, an image, or both. It returns 4096-dim vectors that can be truncated, and shares the text embedding space of Nemotron 3 Embed 8B. NVIDIA: Nemotron 3.5 Rerank 8B (free) Nemotron 3.5 Rerank 8B (free) 20.6M tokens by nvidia Oct 21, 2026 $0 /M input tokens $0 /M output tokens Mistral: Voxtral TTS 3 Voxtral TTS 3 1.23M tokens Voxtral TTS 3 is Mistral's text-to-speech model. It converts text into natural-sounding speech, offers a library of preset voices, supports zero-shot voice cloning, and accepts emotion and pace tags in the input to control delivery. It streams audio with low time-to-first-audio and suits voice agents, narration, and multilingual assistants. by mistralai Oct 20, 2026 $13/M characters Voxtral TTS 3 is Mistral's text-to-speech model. It converts text into natural-sounding speech, offers a library of preset voices, supports zero-shot voice cloning, and accepts emotion and pace tags in the input to control delivery. It streams audio with low time-to-first-audio and suits voice agents, narration, and multilingual assistants. Mistral: Voxtral Small Transcribe 3 Voxtral Small Transcribe 3 15M characters Voxtral Small Transcribe 3 is Mistral's flagship speech-to-text model. It takes audio input and returns a transcript with timestamps and speaker diarization, and supports context biasing for domain-specific vocabulary. Suited for high-accuracy multilingual transcription of meetings, interviews, media, and support calls. by mistralai Oct 20, 2026 $0.000067/second Voxtral Small Transcribe 3 is Mistral's flagship speech-to-text model. It takes audio input and returns a transcript with timestamps and speaker diarization, and supports context biasing for domain-specific vocabulary. Suited for high-accuracy multilingual transcription of meetings, interviews, media, and support calls. Mistral: Voxtral Mini Transcribe 3 Voxtral Mini Transcribe 3 3M characters Voxtral Mini Transcribe 3 is Mistral's compact speech-to-text model. It takes audio input and returns a transcript with timestamps and speaker diarization. It also supports context biasing for domain terms. It is small enough to self-deploy in resource-constrained environments, and fits meeting notes, call analytics, and voice-agent pipelines where cost matters. by mistralai Oct 20, 2026 $0.000058/second Voxtral Mini Transcribe 3 is Mistral's compact speech-to-text model. It takes audio input and returns a transcript with timestamps and speaker diarization. It also supports context biasing for domain terms. It is small enough to self-deploy in resource-constrained environments, and fits meeting notes, call analytics, and voice-agent pipelines where cost matters. MiniMax: MiniMax M3.1 Flash MiniMax M3.1 Flash 1.91B tokens MiniMax-M3.1-Flash is a multimodal model from MiniMax. Compared with M3, it delivers significant improvements across multiple benchmarks, particularly in complex agentic tasks, coding, and cybersecurity. It is built for complex, multi-step co-work automation, understanding real codebases, fixing bugs, and building features, and visual understanding and creation, including frontend and presentation creation. Open weights are planned. by minimax Oct 19, 2026 1M context $0.15 /M input tokens $0.60 /M output tokens

MiniMax-M3.1-Flash is a multimodal model from MiniMax. Compared with M3, it delivers significant improvements across multiple benchmarks, particularly in complex agentic tasks, coding, and cybersecurity. It is built for complex, multi-step co-work automation, understanding real codebases, fixing bugs, and building features, and visual understanding and creation, including frontend and presentation creation. Open weights are planned. DeepSeek: DeepSeek V4.1 Flash 1017 DeepSeek V4.1 Flash 1017 19.7T tokens Academia (#6) Finance (#3) Health (#18) Legal (#11) Marketing (#20) + 7 categories DeepSeek V4.1 Flash 1017 is the October 17, 2026 snapshot of DeepSeek V4.1 Flash, served by DeepSeek under the deepseek-flash API model name. V4.1 Flash is a sparse mixture-of-experts model from DeepSeek, and the first built on the company's Causal Encoder-Decoder (CED) architecture. It activates 8B parameters on input and 16B on output from a 552B-parameter backbone, an asymmetric split that keeps per-token compute low relative to the model's total size. Image understanding is native to the architecture, with visual and text embeddings trained jointly from the start of pre-training rather than added afterward as in the earlier experimental V4 Flash Vision Exp . It is suited for coding, terminal, and computer-use agents, along with long-horizon tasks that must run to completion across many steps and long-context analysis. Compressed KV caching cuts cache memory to roughly a quarter of the previous Flash generation, significantly reducing costs on agentic workloads. DeepSeek positions it as the cost-efficient tier of the V4.1 family and reports that it exceeds V4 Pro on performance, speed, and task completion time. by deepseek Oct 17, 2026 1.05M context $0.07 /M input tokens $0.50 /M output tokens DeepSeek V4.1 Flash 1017 is the October 17, 2026 snapshot of DeepSeek V4.1 Flash, served by DeepSeek under the deepseek-flash API model name. V4.1 Flash is a sparse mixture-of-experts model from DeepSeek, and the first built on the company's Causal Encoder-Decoder (CED) architecture. It activates 8B parameters on input and 16B on output from a 552B-parameter backbone, an asymmetric split that keeps per-token compute low relative to the model's total size. Image understanding is native to the architecture, with visual and text embeddings trained jointly from the start of pre-training rather than added afterward as in the earlier experimental V4 Flash Vision Exp . It is suited for coding, terminal, and computer-use agents, along with long-horizon tasks that must run to completion across many steps and long-context analysis. Compressed KV caching cuts cache memory to roughly a quarter of the previous Flash generation, significantly reducing costs on agentic workloads. DeepSeek positions it as the cost-efficient tier of the V4.1 family and reports that it exceeds V4 Pro on performance, speed, and task completion time. Elorian: Glimpse Glimpse 3.32B tokens Glimpse is a multimodal visual reasoning model from Elorian, built for visual understanding and complex visual reasoning in the physical world. It reasons directly over images and video rather than converting them to text first, so it can find small details in dense scenes, track events across video frames, and ground its answers in what is actually on screen. It is suited for agentic visual workflows such as visual search, video captioning, and dense counting, and for physical-world use cases in robotics, construction, engineering, and manufacturing. by elorian Oct 17, 2026 262K context $0.60 /M input tokens $3 /M output tokens Glimpse is a multimodal visual reasoning model from Elorian, built for visual understanding and complex visual reasoning in the physical world. It reasons directly over images and video rather than converting them to text first, so it can find small details in dense scenes, track events across video frames, and ground its answers in what is actually on screen. It is suited for agentic visual workflows such as visual search, video captioning, and dense counting, and for physical-world use cases in robotics, construction, engineering, and manufacturing. Qwen: Qwen Image 2.1 Turbo Qwen Image 2.1 Turbo 95.5M tokens Qwen Image 2.1 Turbo is the fast, lower-cost tier of the Qwen Image 2.1 series from Qwen, for image generation and editing where throughput and price matter more than peak fidelity. See Qwen Image 2.1 Pro for the higher-quality tier. by qwen Oct 16, 2026 from $0.0184/image Qwen Image 2.1 Turbo is the fast, lower-cost tier of the Qwen Image 2.1 series from Qwen, for image generation and editing where throughput and price matter more than peak fidelity. See Qwen Image 2.1 Pro for the higher-quality tier. Qwen: Qwen Image 2.1 Pro Qwen Image 2.1 Pro 52.1M tokens Qwen Image 2.1 Pro is an image generation and editing model from Qwen, the higher-quality tier of the Qwen Image 2.1 series. Choose it over Qwen Image 2.1 Turbo when output fidelity matters more than cost or speed. by qwen Oct 16, 2026 from $0.046/image Qwen Image 2.1 Pro is an image generation and editing model from Qwen, the higher-quality tier of the Qwen Image 2.1 series. Choose it over Qwen Image 2.1 Turbo when output fidelity matters more than cost or speed. Perplexity: Decider Flash V1.2 Decider Flash V1.2 15B tokens Trivia (#38) Decider Flash V1.2 is a 4B-parameter decision model from Perplexity, a smaller, lower-cost sibling of Decider V1.2 with the same API contract. Instead of generating text, it reads content passed as state (text, JSON, or images) and returns typed, probabilistic answers to one or more named questions in a single request: the probability of yes for a yes/no question ( noul ), a probability for every option plus the most likely one ( choice , up to 255 options), or a probability for every level of an ordered rubric plus the expected score ( score ). It suits high-volume classification, routing, and moderation where application code thresholds the returned numbers rather than parsing a chat reply. by perplexity Oct 16, 2026 $0.01 /M input tokens $0 /M output tokens Decider Flash V1.2 is a 4B-parameter decision model from Perplexity, a smaller, lower-cost sibling of Decider V1.2 with the same API contract. Instead of generating text, it reads content passed as state (text, JSON, or images) and returns typed, probabilistic answers to one or more named questions in a single request: the probability of yes for a yes/no question ( noul ), a probability for every option plus the most likely one ( choice , up to 255 options), or a probability for every level of an ordered rubric plus the expected score ( score ). It suits high-volume classification, routing, and moderation where application code thresholds the returned numbers rather than parsing a chat reply. Perplexity: Decider V1.2 Decider V1.2 101B tokens Legal (#40) Marketing (#37) Translation (#35) Trivia (#3) Decider V1.2 is a new 27B checkpoint of Perplexity's decision model, succeeding Decider V1.1 27B with the same API contract. Instead of generating text, it reads content passed as state (text, JSON, or images) and returns typed, probabilistic answers to one or more named questions in a single request: the probability of yes for a yes/no question ( noul ), a probability for every option plus the most likely one ( choice , up to 255 options), or a probability for every level of an ordered rubric plus the expected score ( score ). It is built for classification, routing, moderation, and rubric grading where application code thresholds the returned numbers rather than parsing a chat reply. For a smaller, lower-cost option with the same contract, see Decider Flash V1.2 . by perplexity Oct 16, 2026 $0.02 /M input tokens $0 /M output tokens Decider V1.2 is a new 27B checkpoint of Perplexity's decision model, succeeding Decider V1.1 27B with the same API contract. Instead of generating text, it reads content passed as state (text, JSON, or images) and returns typed, probabilistic answers to one or more named questions in a single request: the probability of yes for a yes/no question ( noul ), a probability for every option plus the most likely one ( choice , up to 255 options), or a probability for every level of an ordered rubric plus the expected score ( score ). It is built for classification, routing, moderation, and rubric grading where application code thresholds the returned numbers rather than parsing a chat reply. For a smaller, lower-cost option with the same contract, see Decider Flash V1.2 . XingChen: Xing4.0 29B A4B (free) Xing4.0 29B A4B (free) 49.2B tokens Xing4.0-29B-A4B is an open-weight mixture-of-experts language model from China Telecom's AI team (TeleAI) and the successor to the TeleChat series. It has 29B total parameters with 4B active per token and a native 256K-token context window. It was trained entirely on Ascend NPUs with the MindSpore framework and is tuned for complex engineering and agent work: tool calling, multi-step planning, and long-context execution. Reasoning is on by default and can be turned off. by xingchen-agi Oct 16, 2026 262K context $0 /M input tokens $0 /M output tokens Xing4.0-29B-A4B is an open-weight mixture-of-experts language model from China Telecom's AI team (TeleAI) and the successor to the TeleChat series. It has 29B total parameters with 4B active per token and a native 256K-token context window. It was trained entirely on Ascend NPUs with the MindSpore framework and is tuned for complex engineering and agent work: tool calling, multi-step planning, and long-context execution. Reasoning is on by default and can be turned off. Perplexity: Decider Flash Latest Decider Flash Latest This model always redirects to the latest model in the Decider Flash family. by perplexity Oct 16, 2026 $0.01 /M input tokens $0 /M output tokens This model always redirects to the latest model in the Decider Flash family. Reflection: Beam 1.0 501B A23B Beam 1.0 501B A23B 118B tokens Beam 1.0 is Reflection's open-weight reasoning model for coding and long-horizon agentic workflows. It uses a sparse mixture-of-experts architecture with 501B total and 23B active parameters per token, and a 1M-token context window. Beam targets software engineering, multi-step tool use, and mathematical and scientific reasoning, with an emphasis on token efficiency. It takes text input and produces text output, supports configurable reasoning effort, function calling and structured outputs, and is released under Apache 2.0. by reflection Oct 16, 2026 1.05M context $0.25 /M input tokens $1 /M output tokens Beam 1.0 is Reflection's open-weight reasoning model for coding and long-horizon agentic workflows. It uses a sparse mixture-of-experts architecture with 501B total and 23B active parameters per token, and a 1M-token context window. Beam targets software engineering, multi-step tool use, and mathematical and scientific reasoning, with an emphasis on token efficiency. It takes text input and produces text output, supports configurable reasoning effort, function calling and structured outputs, and is released under Apache 2.0.