Text & Chat Models
Powerful language models for conversation, analysis, and content generation
Text and chat models for production AI workloads
Large language models are the workhorse of modern AI: chatbots, agents, summarisers, classifiers, translators. The category is the most crowded on Railwail — OpenAI, Anthropic, Google, Mistral, Meta, DeepSeek, xAI, and dozens of open-weights labs all compete here.
45 models available
Claude Opus 4
Anthropic's most powerful model. Exceptional at complex analysis, agentic tasks, and extended reasoning.
Claude Sonnet 4
Anthropic's most capable model. Excellent for complex analysis, coding, math, and creative writing.
DeepSeek V3.1
DeepSeek's refreshed V3.1 release. 671B MoE / 37B active. Tops open-weights leaderboards on coding and reasoning.
DeepSeek V4 Pro
DeepSeek's April 2026 flagship. 1.6T MoE / 49B active params, 1M context, rivals top closed-source models on STEM and coding at a fraction of the price.
Gemini 2.0 Flash
Google's fastest multimodal model. Supports text, images, audio, and video input.
Gemini 2.5 Pro
Google's latest thinking model. Excels at reasoning, coding, math, and science with massive context window.
GPT-4.1
OpenAI's newest flagship model. Improved reasoning, instruction following, and coding over GPT-4o.
GPT-4o
OpenAI's most capable multimodal model. Excellent for complex reasoning, coding, and creative tasks.
Grok 4
xAI's flagship reasoning model with vision and tool use. 256k context, strong at complex reasoning and STEM tasks.
Kimi K2 (Moonshot)
Moonshot AI's 1T-parameter MoE model. Industry-leading agentic coding and tool-use benchmarks.
MiniMax-01
MiniMax's 456B hybrid lightning-attention model with native 4M-token context. Industry-leading long-context.
o3-mini
OpenAI's reasoning model optimized for STEM tasks, coding, and math. Uses chain-of-thought reasoning.
Perplexity Sonar Pro
Perplexity's premium web-grounded search model with multi-step reasoning over live sources.
Qwen 3 235B Instruct
Alibaba's Qwen 3 flagship MoE: 235B total / 22B active. Strong reasoning and tool use, open-weights.
AI21 Jamba 1.5 Large
AI21's flagship hybrid Mamba-Transformer model with a 256k context window for long-document tasks.
AI21 Jamba 1.5 Mini
Cost-efficient hybrid Mamba-Transformer model with 256k context. Tuned for high-throughput RAG.
Claude Haiku 3.5
Anthropic's fast and affordable model. Great for quick tasks, summarization, and simple coding.
Cohere Aya 23 35B
Open-weights multilingual research model from Cohere covering 23 languages. 35B parameters.
Cohere Command Light (legacy)
Cohere's fast lightweight chat model (deprecated Sep 2025). Kept as comparison tombstone.
Cohere Command R (08-2024)
Cohere's mid-tier RAG/tool model. Cost-efficient sibling of Command R+ with 128k context.
Cohere Command R+ (08-2024)
Cohere's flagship RAG- and tool-optimized chat model. 128k context, refreshed August 2024.
DeepSeek R1
DeepSeek's reasoning model with chain-of-thought capabilities. Excellent for complex problem-solving.
DeepSeek V3
Powerful open-weight model from DeepSeek. Strong at coding, math, and Chinese/English tasks.
DeepSeek V4 Flash
Efficiency-optimized variant of DeepSeek V4. 284B MoE / 13B active, 1M context, ultra-low pricing for high-throughput workloads.
GPT-4o Mini
Small, fast, and affordable model for lightweight tasks. Great balance of speed and capability.
Grok 3
xAI's flagship model. Strong at reasoning, coding, and real-time knowledge with web search capabilities.
Llama 3.3 70B
Meta's open-source 70B parameter model. Strong all-around performance with multilingual support.
M2M-100 12B
Meta M2M-100 12B many-to-many translation model. Direct translation between 100 languages without pivoting through English.
MADLAD-400 3B
Google MADLAD-400 3B multilingual translation model. 419 languages supported, trained on a 5T-token multilingual corpus with strong low-resource performance.
mBART 50 Many-to-Many
Meta mBART-50 many-to-many translation model. 50 supported languages with strong performance on news and conversational text.
Microsoft Phi-3.5 MoE Instruct
Mixture-of-experts Phi-3.5: 42B total / 6.6B active params. 128k context, multilingual.
Mistral Large
Mistral's flagship model. Strong reasoning, multilingual, and coding capabilities.
NLLB-200 3B
Meta's No Language Left Behind 3.3B translation model. Direct translation between any pair of 200+ languages including many low-resource African and Asian languages.
NLLB-200 Distilled 600M
Meta's distilled 600M NLLB. Same 200-language coverage as the 3B model with a fraction of the parameters, ideal for edge or high-throughput deployment.
Nous Hermes 3 405B
Full-parameter fine-tune of Llama 3.1 405B by Nous Research. Steerable, uncensored, strong tool use.
Nous Hermes 3 70B
Llama-3.1-70B fine-tune from Nous Research with strong tool/agent capabilities and uncensored alignment.
Perplexity Sonar
Perplexity's fastest and cheapest web-grounded chat model. Live-source citations included.
Perplexity Sonar Reasoning
Perplexity's reasoning model with chain-of-thought and integrated web search.
Qwen 2.5 72B
Alibaba's powerful open-source model. Excellent at coding, math, and multilingual tasks.
Qwen 2.5-Max
Alibaba's flagship pretrained MoE model. Top-tier reasoning and code performance via DashScope API.
SeamlessM4T v2 Large (Text)
Meta SeamlessM4T v2 Large. Universal multilingual translation across 100+ languages with text-to-text mode for documents and chat.
Snowflake Arctic Instruct
Snowflake's open MoE model: 480B total / 17B active params with dense+MoE hybrid architecture.
TII Falcon 180B Chat
TII's 180B causal decoder chat model fine-tuned on Ultrachat, Platypus and Airoboros.
TowerInstruct 13B
Unbabel TowerInstruct 13B. Llama-2-based multilingual translation and post-editing model. Strong terminology consistency for enterprise localization.
Yi Large
01.AI's larger general-purpose chat model with 32k context window and strong bilingual performance.
Top text & chat models picks
Hand-picked across four common criteria — resolved against the live catalog so the picks track price and performance changes.
Anthropic's most powerful model. Exceptional at complex analysis, agentic tasks, and extended reasoning.
Learn moreCohere's mid-tier RAG/tool model. Cost-efficient sibling of Command R+ with 128k context.
Learn moreMiniMax's 456B hybrid lightning-attention model with native 4M-token context. Industry-leading long-context.
Learn moreSmall, fast, and affordable model for lightweight tasks. Great balance of speed and capability.
Learn morePricing in text generation is almost always per-token, split into a cheaper input rate and a more expensive output rate. A million input tokens is roughly 750,000 English words, so even verbose prompts rarely push the per-call price above a few cents. Output is where bills grow: agentic loops that re-prompt themselves, long reports, and uncached few-shot examples add up quickly. Cache the system prompt, trim history aggressively, and prefer JSON-mode for structured tasks to keep token counts down.
The trade-off triangle is quality, latency, and cost. Flagship models (GPT-5, Claude 4.6, Gemini 2.5) cost ten to fifty times more than budget tiers (Haiku, Flash, Mini) and respond two to four times slower, but they reason deeper, follow instructions more reliably, and hallucinate less. For high-volume classification or extraction, a budget model is almost always the right call. For long-form analysis, code review, or anything user-facing, the flagship usually pays for itself.
Watch out for context dilution: when you stuff 200K tokens into the window, the model's attention spreads thin and it starts to ignore the middle of the prompt — even on long-context flagships. Retrieve the relevant 8-16K tokens with embeddings instead of pasting the whole document.
Popular use cases
Common patterns built with text & chat models on Railwail.
Frequently asked questions
Start Building with AI
Access all models through a single API. Get free credits when you sign up — no credit card required.