How NeverMetered compares
Price, model, usage caps, parallel streams and context for coding-agent inference options.
Data checked as of 18 September 2026. Providers change prices and limits often — verify on their own site before deciding.
| Plan | Price | Model | Caps | Streams | Context | Verified |
|---|---|---|---|---|---|---|
| NeverMetered — Day Pass | $2.99 for 1 day | Qwen3.8-27B | No token quota, no 5-hour window and no weekly cap | 1 | 128K | Live pricing |
| NeverMetered — Week Pass | $11.99 for 7 days | Qwen3.8-27B | No token quota, no 5-hour window and no weekly cap | 1 | 128K | Live pricing |
| NeverMetered — Solo | $29 / month | Qwen3.8-27B | No token quota, no 5-hour window and no weekly cap | 1 | 256K | Live pricing |
| NeverMetered — Pro | $79 / month | Qwen3.8-27B | No token quota, no 5-hour window and no weekly cap | 3 | 256K | Live pricing |
| Z.ai GLM Coding Plan | From $18/mo (Lite tier; Pro and Max cost more) | GLM-5.3 and GLM-5.3-Flash (legacy GLM-5.2/5.1/4.7 requests auto-route to the current models) | Prompt-credit quota on a rolling 5-hour window plus a weekly cap (e.g. Lite: 2,000 credits/5hr, 10,000/week; Max: 28,000/5hr, 140,000/week), with credits weighted by input/cached/output token type | Not stated | Not stated | Checked 18 September 2026 |
| OpenCode Zen | Pay-as-you-go, no flat monthly fee ($20 starting balance, auto-reloads at $5) | 50+ hosted models across providers (Claude, GPT, Gemini, Qwen, DeepSeek, GLM, Kimi, MiniMax), plus some free-tier models | No fixed quota; billed per token consumed, with an optional configurable monthly spend limit per workspace | Not stated | Not stated (varies by model) | Checked 18 September 2026 |
| Synthetic | $30/mo (or $1/day) per pack | DeepSeek-V4.1-Flash, Kimi-K3, NVIDIA Nemotron-3 Super, GPT-OSS-120B, Qwen3.8-27B, GLM-4.7-Flash, GLM-5.3-Flash | 500 requests per 5-hour window per pack, with credits that regenerate over time; buying extra packs adds capacity | 1 concurrent request per model per pack (additional packs add concurrency) | 128K-512K depending on model (e.g. GLM-5.3-Flash and Kimi-K3 at 512K, GPT-OSS-120B at 128K) | Checked 18 September 2026 |
| ArliAI | $10-$160/mo across 5 tiers (Starter to Max) | Open-source LLMs up to 355B parameters; the specific model ceiling scales with the plan tier | No per-token metering inside a tier; caps instead come from a maximum context window and a request-concurrency limit, both of which scale up by tier | 1 request at a time on the three lower tiers, 2 parallel on Pro, up to 6 parallel on the top Max tier | 16K (Starter) up to 512K (Max), scaling with tier | Checked 18 September 2026 |
| Claude Max (Anthropic) | From $100/mo (5x usage tier); a 20x tier is also offered at a higher, unpublished-on-page price | Claude Opus, Sonnet, Haiku, and Fable | Usage resets on a rolling 5-hour session window with an added weekly cap on top; Anthropic states it may apply further caps (e.g. monthly, feature-specific) at its discretion | Not stated | Up to 1M tokens (varies by model) | Checked 18 September 2026 |
| ChatGPT Pro (Codex) | From $100/mo | GPT-6 Astra (Pro reasoning) plus the GPT-5.6 series (Sol/Terra/Luna); Codex tasks run on these | "5x more usage" than Plus, with "maximum Codex tasks" subject to abuse guardrails rather than a published hard number | Not stated | 128K (instant) / 400K (reasoning) total context window | Checked 18 September 2026 |