Skip to content

State of Open Weights — the open models we serve, and why

Living report · updated monthly (numbers) · on lineup changes (roster) · last updated

Open a model router today and you’ll find 400+ models — the same family in five quantizations, three deprecation states, a dozen providers with different latencies and silent fallbacks. Choice isn’t the product; it’s a tax. This report is the opposite: the small set of open-weight models we actually operate, each one there for a job, with the facts that matter — real license terms, architecture, context, current prices and scores — kept current so you never have to reconcile stale numbers across blog posts.

Sources: model facts from the official Hugging Face cards (linked per row); prices and Intelligence Index from Artificial Analysis, refreshed monthly by the same collector that updates the Pareto Frontier report.

ModelPoolThe jobParamsContextLicense
Kimi K2.7FrontierAgentic coding, tool use — the newest Kimi1T MoE / 32B active256KModified MIT¹
Kimi K2.6FrontierBest open SWE-bench Verified score; pinned agent workflows1T MoE / 32B active256KModified MIT¹
GLM 5.2FrontierTop-scoring open model on the AA index; long-horizon coding753B MoE1MMIT
MiniMax M3Frontier1M context + native multimodal input~428B MoE / 23B active1MCommunity License²
DeepSeek V4 FlashCoreHigh-volume workloads: extraction, chat, summarization, RAG284B MoE / 13B active1MMIT
MiMo V2.5CoreBudget multimodal; strongest coding card in its price class310B MoE / 15B active1MMIT

¹ Modified MIT: attribution required only above 100M MAU or $20M/mo revenue. ² MiniMax Community License: “Built with MiniMax M3” attribution; >$20M/yr revenue requires written authorization. “Open” is not one thing — MIT and Apache 2.0 mean unrestricted commercial use; community licenses sit between open and proprietary. For API consumers none of this matters day to day; it matters if you later self-host or embed weights in a product.

ModelInput $/MOutput $/MAA IndexHeadline benchmark
Kimi K2.7$0.95$4.0042MCP Mark Verified 81.1 (vendor suite)
Kimi K2.6$0.95$4.0043SWE-bench Verified 80.2 · Pro 58.6
GLM 5.2$1.40$4.4051SWE-bench Pro 62.1 · Terminal-Bench 2.1 81.0
MiniMax M3$0.30$1.2044SWE-bench Verified 80.5 · Pro 59.0
DeepSeek V4 Flash$0.14$0.2840LiveCodeBench 91.6
MiMo V2.5$0.105$0.28SWE-bench Pro 56.1 · Terminal-Bench 2 65.8

Prices are the vendors’ list prices per million tokens as tracked by Artificial Analysis (MiMo V2.5: OpenRouter listing — no AA entry yet). Benchmark scores are as published on each model’s official card; scaffolding differs between labs. Where these models sit against the closed flagships — and which are Pareto-efficient — is the Pareto Frontier report; tier-fair matchups are in the Which-LLM guide.

Frontier when the work is agentic coding, multi-step agents, or anything where a failed run costs more than the tokens did. Core when the work is volume — extraction, classification, summarization, chat, pipelines that run all day. Both are flat-rate: no token caps during your reserved hours, so the per-token prices above stop mattering once you’re inside your window.

  • 2026-07-06 (first edition) — baseline roster: Kimi K2.7, Kimi K2.6, GLM 5.2, MiniMax M3 (Frontier); DeepSeek V4 Flash, MiMo V2.5 (Core). This report absorbs the per-model tables previously maintained in blog posts, which now link here instead of carrying their own copies of the numbers.

The live pool menu — with current block prices and availability — is always at cheapestinference.com/pools: Frontier from $44.20/mo, Core from $5.94/mo, one OpenAI- and Anthropic-compatible API.