Topic collection
Language Model APIs
Model reviews and comparisons covering context, tool use, coding workflows, pricing, and API access.
The language-model guides compare APIs by the work they need to perform, including structured output, tool use, coding, long-context processing, latency, and price. They avoid treating a single benchmark as a universal ranking. Use the articles to understand tradeoffs and the linked model pages to inspect current capabilities before selecting a model for an application or agent workflow.
9 articles
Articles about language models
Claude Fable 5 vs Kimi K3: API Pricing and Benchmarks (2026)
Kimi K3 runs $2.10/1M input against Fable 5's $5.50, but Fable 5 leads Arena 1507 to 1486 and HLE 53.3 to 43.5. Pricing, benchmarks, and code.
GPT 5.6 Sol vs Kimi K3: API Pricing and Benchmarks (2026)
Sol runs $1.50/1M input against Kimi K3's $2.10 and scores 80 on the Coding Agent Index. Pricing, benchmarks, speed, and code to call both.
Claude API Pricing: $0.30–$5.50 per 1M Input Tokens (2026)
Claude API pricing in 2026: Opus 5 lists at $5/$25 and Sonnet 5 at $2/$10 per 1M tokens. Unifically serves the same models from $0.30 per 1M input tokens.
DeepSeek API: Pricing, Benchmarks, and How to Access It (2026)
Yes, DeepSeek has a public API: V4-Pro lists at $1.32/1M input, $3.96/1M output peak, half off-peak. Release dates, benchmarks, cache pricing, and access.
Gemini 3.7 Flash API: Pricing, Benchmarks, and How to Access It (2026)
Gemini 3.7 Flash: 1M context, ~340 tokens per second, and a nine-point coding jump over 3.6 Flash. Benchmarks, pricing, and API access.
GPT-5.6 Sol Ultrafast API: 750 Tokens per Second on Cerebras (2026)
OpenAI's Ultrafast mode runs GPT-5.6 Sol at up to 750 output tokens per second on Cerebras hardware. How it works, who gets it, and what to call today.
Grok 4.6 API: Benchmarks, Pricing, and How to Access It (2026)
Grok 4.6 scores 95.6% on SWE-bench Verified and takes first on CursorBench 3.2. Model ID xai/grok-4.6, 500k context, benchmarks, and API access.
Qwen3.8 27B: Benchmarks, Specs, and How to Run It (2026)
Qwen3.8 27B is a 27B open-weight VL model under Apache 2.0: 61.7% SWE-bench Pro, 262K context, image and video input. Benchmarks, VRAM math, how to run it.
Kimi K3 API: Pricing, Benchmarks, and How to Access It (2026)
Kimi K3 API costs $2.10/1M input and $10.50/1M output on Unifically. 2.8T parameters, 1M-token context, benchmarks vs Claude and GPT, weights due July 27.