fix(free-tiers): remove unsupported Bytez and Nous free claims

The shipped catalog classified Bytez 70B/72B models and two Nous Hermes-4 models as free without evidence that their free plans cover those models. Because `isFreeModel` trusts the shipped catalog, these entries could populate free-only imports even when the provider payload reported paid pricing.

Remove those four free-catalog entries while preserving the provider/model registries. Keep Bytez Mistral-7B as recurring credit, with zero token counters representing an unpublished token equivalent instead of the unsupported 1M estimate. Refresh the evidence comments, curation date and relevant documentation. Measured catalog totals: 485 entries, 53 eligible recurring/keyless providers and 35 recurring pools; rounded 1.62B/2.22B headlines are unchanged. Diagram and generated screenshot counts are aligned.

Evidence checked on 2026-10-15: [Bytez billing](https://docs.bytez.com/model-api/docs/billing) limits free open models to 7B and requires purchased-credit unlocks for the larger tiers; [Nous Portal](https://portal.nousresearch.com/) does not establish free coverage for the named Hermes-4 models. This does not assert that Nous has no free models.

Reconciled by a normal merge of release commit `9dae75f291cb9ac1dc0862a0ab46879df89687d8`. The only conflict was the curation date; the later 2026-10-15 date is retained alongside all SambaNova catalog and screenshot changes from #16117.

Validation with exact lockfile dependencies:

- RED before production edits: eight regression tests, seven failures and one passing control.
- Integrated focused tests: 13/13, including both SambaNova regressions. Full native importer set: 52 files, 394/394 tests. Vitest importer set: five files, 45/45 tests. No skipped/cancelled tests.
- Formatting, scoped ESLint, file-size, changelog integrity, test discovery, dashboard baseline gate, cyclomatic/cognitive complexity and full docs-all passed. Dashboard retains 195 existing errors within its frozen baseline; complexity is 3045/3218 and cognitive 1404/1437. The generated screenshot matches the candidate byte-for-byte.
- `typecheck:core` remains red: all 57 diagnostic signatures reproduce identically on the untouched exact base SHA, in three files byte-identical to that base (`request/openai-to-claude.ts`, `response/openai-to-claude.ts`, `utils/dsmlToolCalls.ts`). This is an inherited failure tracked by #15306, not a passing gate. The release baseline was not widened.

The earlier mixed-run runner error and documentation-count correction are preserved in the local evidence; the integrated importer runs above were executed cleanly with their correct runners. A live provider benchmark was not run and is not claimed as passing. Normal commit/push hooks are retained.

Closes #16781.

⚠️ base-red inherited: #15306

Thanks @diegosouzapw.
This commit is contained in:
Diego Rodrigues de Sa e Souza
2026-10-15 23:18:57 -03:00
committed by GitHub
parent c17f596e64
commit 6aec22c1ca
9 changed files with 95 additions and 34 deletions

View File

@@ -19,9 +19,9 @@
> **Bring your own provider accounts.** This is an estimated aggregate across separately eligible third-party free tiers, not a token grant from OmniRoute. Sign up, obtain credentials where required, and connect the providers you can use; each provider controls its limits, availability, and terms.
>
> Stacking free tiers by hand is painful — dozens of SDKs, dozens of rate limits, and no idea how much you actually have. OmniRoute catalogs **489 free-tier entries across 35 recurring pool keys** and computes the token headline from the **17 pools with a published positive monthly budget plus five per-model Groq caps**, deduplicated by shared pool. Quotas that only open after a regional identity check (today: ModelScope) are shown apart, +~6M behind regional identity verification, and never summed into the headline. The result stays visible on the dashboard (`/dashboard/free-tiers`).
> Stacking free tiers by hand is painful — dozens of SDKs, dozens of rate limits, and no idea how much you actually have. OmniRoute catalogs **485 free-tier entries across 35 recurring pool keys** and computes the token headline from the **17 pools with a published positive monthly budget plus five per-model Groq caps**, deduplicated by shared pool. Quotas that only open after a regional identity check (today: ModelScope) are shown apart, +~6M behind regional identity verification, and never summed into the headline. The result stays visible on the dashboard (`/dashboard/free-tiers`).
<img src="./docs/diagrams/free-tier-budget.svg" width="100%" alt="OmniRoute free-tier budget card: ~1.62B free tokens per month steady, up to ~2.22B in the first month with signup credits, from 35 documented recurring pool keys covering 489 cataloged free-tier entries behind one endpoint. Honest pool-deduped math — each shared pool counted once, including 17 recurring pools with a published positive monthly token budget plus five per-model Groq caps; 13 providers are marked avoid in the terms-risk catalog so you decide. Budget bar includes Mistral 1B, Nara 210M, LLM7 150M, xKiro 150M, Groq 30M (five per-model caps) and smaller pools, plus first-month signup credits and permanently-free no-token-cap providers surfaced separately so they never inflate the headline. Live used/remaining on /dashboard/free-tiers."/>
<img src="./docs/diagrams/free-tier-budget.svg" width="100%" alt="OmniRoute free-tier budget card: ~1.62B free tokens per month steady, up to ~2.22B in the first month with signup credits, from 35 documented recurring pool keys covering 485 cataloged free-tier entries behind one endpoint. Honest pool-deduped math — each shared pool counted once, including 17 recurring pools with a published positive monthly token budget plus five per-model Groq caps; 13 providers are marked avoid in the terms-risk catalog so you decide. Budget bar includes Mistral 1B, Nara 210M, LLM7 150M, xKiro 150M, Groq 30M (five per-model caps) and smaller pools, plus first-month signup credits and permanently-free no-token-cap providers surfaced separately so they never inflate the headline. Live used/remaining on /dashboard/free-tiers."/>
> Animated summary of the live `/dashboard/free-tiers` page. Full methodology (pool dedupe, credit tiers, provider terms): **[docs/reference/FREE_TIERS.md](docs/reference/FREE_TIERS.md)**.
>
@@ -239,7 +239,7 @@ curl http://localhost:20128/v1/chat/completions \
</div>
<img src="./docs/diagrams/promise-pillars.svg" width="100%" alt="The Promise — One endpoint and 372 providers. Automatic fallback keeps routing while another healthy target is available. Six pillars: resilient fallback across 372 providers · up to 95% token savings on eligible workloads · $0 to start with 150+ free tiers and 54 recurring/keyless free-forever providers · 36 CLI/agent integrations through one config · OpenAI, Claude, Gemini and Responses API compatibility at /v1 · production controls including circuit breakers, TLS stealth, MCP 110 tools, A2A, memory, guardrails, evals and 39,000+ static test declarations across 5,100+ tracked test files."/>
<img src="./docs/diagrams/promise-pillars.svg" width="100%" alt="The Promise — One endpoint and 372 providers. Automatic fallback keeps routing while another healthy target is available. Six pillars: resilient fallback across 372 providers · up to 95% token savings on eligible workloads · $0 to start with 150+ free tiers and 53 recurring/keyless free-forever providers · 36 CLI/agent integrations through one config · OpenAI, Claude, Gemini and Responses API compatibility at /v1 · production controls including circuit breakers, TLS stealth, MCP 110 tools, A2A, memory, guardrails, evals and 39,000+ static test declarations across 5,100+ tracked test files."/>
<br/>
<br/>
@@ -680,7 +680,7 @@ of your shell history. → [CLI Integrations](docs/guides/CLI-INTEGRATIONS.md)
</div>
> **372 registered providers** across the canonical chat, media, search, local, cloud-agent and system collections, including **154 carrying `hasFree: true` discovery metadata**. The chat model registry covers **237 providers / 3,009 distinct provider-model pairs / 1,443 raw model IDs**; the separate free-budget catalog has **491 per-model rows**, **35 recurring pools** and **54 recurring/keyless free-forever providers**. These are different denominators by design; definitions and pool-deduped calculations live in the [Provider Reference](docs/reference/PROVIDER_REFERENCE.md) and [Free Tiers](docs/reference/FREE_TIERS.md).
> **372 registered providers** across the canonical chat, media, search, local, cloud-agent and system collections, including **154 carrying `hasFree: true` discovery metadata**. The chat model registry covers **237 providers / 3,009 distinct provider-model pairs / 1,443 raw model IDs**; the separate free-budget catalog has **485 per-model rows**, **35 recurring pools** and **53 recurring/keyless free-forever providers**. These are different denominators by design; definitions and pool-deduped calculations live in the [Provider Reference](docs/reference/PROVIDER_REFERENCE.md) and [Free Tiers](docs/reference/FREE_TIERS.md).
<div align="center">
@@ -1359,7 +1359,7 @@ Métricas canônicas em 2026-08-24: **1.029 vídeos únicos** · **11.132.922 vi
<tr><td nowrap><b><a href="docs/architecture/RESILIENCE_GUIDE.md">Resilience Guide</a></b></td><td>Circuit breakers, cooldowns, queue, anti-thundering herd, TLS spoofing</td></tr>
<tr><td nowrap><b><a href="docs/routing/AUTO-COMBO.md">Auto-Combo Engine</a></b></td><td>16-factor scoring, mode packs, self-healing</td></tr>
<tr><td nowrap><b><a href="docs/ops/PROXY_GUIDE.md">Proxy Guide</a></b></td><td>3-level proxy system, 1proxy marketplace, registry CRUD</td></tr>
<tr><td nowrap><b><a href="docs/reference/FREE_TIERS.md">Free Tiers</a></b></td><td>Consolidated directory: 35 documented recurring pools / 489 cataloged free-tier entries</td></tr>
<tr><td nowrap><b><a href="docs/reference/FREE_TIERS.md">Free Tiers</a></b></td><td>Consolidated directory: 35 documented recurring pools / 485 cataloged free-tier entries</td></tr>
<tr><td nowrap><b><a href="docs/guides/FEATURES.md">Features Gallery</a></b></td><td>Visual dashboard tour with screenshots</td></tr>
<tr><td nowrap><b><a href="docs/architecture/CODEBASE_DOCUMENTATION.md">Codebase Documentation</a></b></td><td>Beginner-friendly codebase walkthrough</td></tr>
</table>

View File

@@ -0,0 +1 @@
- Correct free-model classification for Bytez paid 70B/72B tiers and unverified Nous Hermes-4 entries; retain Bytez Mistral-7B without an unsupported token allowance and refresh the documented catalog counts. Closes #16781.

View File

@@ -64,7 +64,7 @@
<text x="60" y="228" font-family="Consolas, 'Courier New', monospace" font-size="104" font-weight="800" fill="url(#gradBrandF)">~1.62B</text>
<text x="62" y="266" font-family="Consolas, 'Courier New', monospace" font-size="15" letter-spacing="3" font-weight="700" fill="#a1a1aa">FREE TOKENS / MONTH &#183; <tspan fill="#22c55e">STEADY</tspan></text>
<text x="62" y="298" font-family="Inter, 'Segoe UI', Arial, Helvetica, system-ui, sans-serif" font-size="16" fill="#F7F6FC">up to <tspan font-weight="800" fill="#22c55e">~2.22B</tspan> in your first month &#8212; signup credits</text>
<text x="62" y="326" font-family="Consolas, 'Courier New', monospace" font-size="12" fill="#71717a">documented free tiers &#183; <tspan fill="#8b5cf6">35 recurring pools</tspan> &#183; <tspan fill="#8b5cf6">489 catalog entries</tspan> &#183; one endpoint</text>
<text x="62" y="326" font-family="Consolas, 'Courier New', monospace" font-size="12" fill="#71717a">documented free tiers &#183; <tspan fill="#8b5cf6">35 recurring pools</tspan> &#183; <tspan fill="#8b5cf6">485 catalog entries</tspan> &#183; one endpoint</text>
<!-- ═══ Panel · The honest math ═══ -->
<rect x="680" y="84" width="460" height="216" rx="14" fill="#161b22" stroke="#ffffff" stroke-opacity="0.08" stroke-width="1"/>

Before

Width:  |  Height:  |  Size: 18 KiB

After

Width:  |  Height:  |  Size: 18 KiB

View File

@@ -1,4 +1,4 @@
<svg viewBox="0 0 1200 540" xmlns="http://www.w3.org/2000/svg" role="img" aria-label="The OmniRoute promise: one endpoint and 372 providers. Six pillars. Resilient fallback: automatic routing continues while another healthy target is available. Save up to 95 percent of eligible tokens: RTK plus Caveman stacked compression averages about 89 percent on tool-heavy sessions. Zero dollars to start: 150+ providers with a free tier and 54 recurring or keyless free-forever providers. Every tool works: 36 CLI and agent integration records, including Claude Code, Codex, Cursor, Cline, Copilot and Antigravity, through one config. One endpoint: OpenAI, Claude, Gemini and Responses API translation at /v1. Production controls: circuit breakers, TLS stealth, MCP with 110 tools, A2A, memory, guardrails, evals, and 39,000+ static test declarations across 5,100+ tracked test files.">
<svg viewBox="0 0 1200 540" xmlns="http://www.w3.org/2000/svg" role="img" aria-label="The OmniRoute promise: one endpoint and 372 providers. Six pillars. Resilient fallback: automatic routing continues while another healthy target is available. Save up to 95 percent of eligible tokens: RTK plus Caveman stacked compression averages about 89 percent on tool-heavy sessions. Zero dollars to start: 150+ providers with a free tier and 53 recurring or keyless free-forever providers. Every tool works: 36 CLI and agent integration records, including Claude Code, Codex, Cursor, Cline, Copilot and Antigravity, through one config. One endpoint: OpenAI, Claude, Gemini and Responses API translation at /v1. Production controls: circuit breakers, TLS stealth, MCP with 110 tools, A2A, memory, guardrails, evals, and 39,000+ static test declarations across 5,100+ tracked test files.">
<desc>Animated promise card: six pillar tiles fade in in reading order, then a soft colored border highlight sweeps from tile to tile in a continuous cycle.</desc>
<defs>
<pattern id="gridPaperP" width="32" height="32" patternUnits="userSpaceOnUse">

Before

Width:  |  Height:  |  Size: 10 KiB

After

Width:  |  Height:  |  Size: 10 KiB

View File

@@ -1,7 +1,7 @@
---
title: "Free Tiers & Free-Token Budget"
version: 3.8.50
lastUpdated: 2026-09-03
lastUpdated: 2026-10-15
---
# Free Tiers & Free-Token Budget
@@ -60,6 +60,16 @@ A 50-agent web-research pass (official docs + last-7-days news, adversarially ve
---
## 2026-10-15 Bytez and Nous correction
[Bytez billing](https://docs.bytez.com/model-api/docs/billing) limits its Free plan's
open models to 7B parameters. The 70B and 72B entries require purchased-credit
unlocks and have been removed. Mistral-7B remains `recurring-credit`: credits
refresh every four weeks, but billing is per inference second and no token
allowance is published. Its token counters are zero to represent an unknown
quantity, replacing the unsupported 1M-token estimate; this does not mean zero
free credits. The Nous changes are detailed in the provider note below.
## Two regimes — counting vs deciding
OmniRoute answers "is it free?" through two regimes that intentionally read
@@ -209,7 +219,7 @@ Most "free tokens per month" figures in this space are sums of per-model labels.
---
## Per-provider free-tier (refreshed 2026-09-02 for the re-audited rows; 2026-06-17 otherwise)
## Per-provider free-tier (Bytez/Nous checked 2026-10-15; re-audited rows 2026-09-02; others 2026-06-17)
> Regenerated from the per-model catalog (`open-sse/config/freeModelCatalog.ts`), pool-deduped. Sorted by recurring steady tokens/mo. `uncapped*` = permanently free but no published token cap (rate/concurrency-limited) — real access, **not** summed into the headline. `—` = credit-only / keyless / not token-quantifiable.
@@ -250,7 +260,7 @@ Most "free tokens per month" figures in this space are sums of per-model labels.
| `deepseek` | signup credit | — | ~5M | ok | 2 |
| `hyperbolic` | signup credit | — | ~5M | ok | 8 |
| `nscale` | signup credit | — | ~5M | caution | 6 |
| `bytez` | signup credit | — | ~1M | ambiguous | 3 |
| `bytez` | credit (rec.) | — | unknown | ambiguous | 1 |
| `deepinfra` | signup credit | — | ~1M | caution | 22 |
| `fireworks` | signup credit | — | ~1M | avoid | 10 |
| `nebius` | signup credit | — | ~1M | caution | 1 |
@@ -270,7 +280,6 @@ Most "free tokens per month" figures in this space are sums of per-model labels.
| `monsterapi` | keyless | — | — | ambiguous | 1 |
| `muse-spark-web` | keyless | — | — | avoid | 3 |
| `nlpcloud` | keyless | — | — | avoid | 1 |
| `nous-research` | keyless | — | — | ambiguous | 2 |
| `nvidia` | keyless | — | — | caution | 13 |
| `opencode` | keyless | — | — | avoid | 7 |
| `pollinations` | keyless | — | — | caution | 31 |
@@ -346,7 +355,7 @@ OpenCode session/UA headers (`opencodeFreeTierContract.ts`) or expect the 403.
- **`muse-spark-web`** — The shipped freeNote ("Free with login — Meta AI platform with Llama models") is broadly accurate regarding login requirement and Llama model access. No tightening of the free tier was detected; it r…
- **`nlpcloud`** — The shipped freeNote says "Trial credits for new accounts," which implies a one-time trial. In reality, NLP Cloud's free tier is a recurring monthly free plan (10,000 requests/month), not trial credi…
- **`nomic`** — Our shipped freeNote says "Free Nomic Embed API" with no qualification, implying ongoing free access. Reality is a one-time 1M-token trial credit only — after that token budget is consumed, paid subs…
- **`nous-research`** — The shipped freeNote ("Free tier: 50 RPM, 500,000 TPM") does not match the current Nous Portal product. The portal launched April 27, 2026 and structures its free tier as $0.10/month in recurring cre…
- **`nous-research`** — Rechecked 2026-10-15: the [Nous Portal Free plan](https://portal.nousresearch.com/) offers free models with $0 monthly credits, but does not establish free access for Hermes-4-405B or Hermes-4-70B. Both unverified entries are excluded from the shipped free catalog. This does not establish that Nous has no free models.
- **`nvidia`** — The "40 RPM, 70+ models" rate limit element matches the catalog, but the freeNote framing as a simple dev-access tier undersells that the old one-time credit pool has been removed — access is now tru…
- **`ollama-cloud`** — Our shipped freeNote is "(none)" — this is stale. Ollama Cloud launched a cloud inference product with a genuine free tier that provides light weekly GPU-time-based access to hosted open models.
- **`openrouter`** — RPD tightened from 200 to 50 for zero-credit accounts (RPM unchanged at 20). The catalog note was accurate on RPM but overstated the RPD by 4x for the no-credits baseline tier. **Runtime tracking (#6842)**: this is no longer just a static note — `open-sse/services/openrouterQuotaFetcher.ts` polls `/api/v1/key` + `/api/v1/credits` for per-key credit cap/remaining/reset and daily/weekly/monthly USD spend, and `open-sse/services/openrouterFreeWindow.ts` locally tracks the `:free`-model 50-or-1000-per-day + 20 RPM windows described above (corrected from `X-RateLimit-*` response headers on 429s), surfaced in Dashboard → Provider Quota.

View File

@@ -3,7 +3,7 @@
<rect x="16" y="16" width="868" height="564" rx="13" fill="#161b22" stroke="#30363d"/>
<text x="868" y="572" fill="#484f58" font-size="10.5" text-anchor="end">OmniRoute · /dashboard/free-tiers · preview mockup</text>
<text x="32" y="50" fill="#e6edf3" font-size="18" font-weight="700">Monthly free-token budget</text>
<text x="868" y="50" fill="#7d8590" font-size="13" text-anchor="end">22 free pools · 489 models · one endpoint</text>
<text x="868" y="50" fill="#7d8590" font-size="13" text-anchor="end">22 free pools · 485 models · one endpoint</text>
<text x="32" y="84" fill="#7d8590" font-size="11.5">Steady / month</text>
<text x="32" y="114" fill="#e6edf3" font-size="27" font-weight="800">~1.62B</text>
<text x="330" y="84" fill="#7d8590" font-size="11.5">First month (+ signup credits)</text>

Before

Width:  |  Height:  |  Size: 8.6 KiB

After

Width:  |  Height:  |  Size: 8.6 KiB

View File

@@ -22,7 +22,7 @@ import type { FreeModelBudget } from "./freeModelCatalog.ts";
* rewrites file timestamps on every deploy, which would report a months-old
* catalog as "updated today". Bump this whenever the entries below change.
*/
export const FREE_CATALOG_CURATED_AT = "2026-10-10";
export const FREE_CATALOG_CURATED_AT = "2026-10-15";
export const FREE_MODEL_BUDGETS: FreeModelBudget[] = [
{ provider: "agentrouter", modelId: "claude-opus-4-8", displayName: "Claude Opus 4.8", monthlyTokens: 0, creditTokens: 200000000, freeType: "one-time-initial", poolKey: "agentrouter", tos: "caution" },
@@ -109,9 +109,11 @@ export const FREE_MODEL_BUDGETS: FreeModelBudget[] = [
{ provider: "bluesminds", modelId: "gemini-2.5-pro", displayName: "Gemini 2.5 Pro (VIP)", monthlyTokens: 7200000, creditTokens: 0, freeType: "recurring-daily", poolKey: "bluesminds", tos: "ambiguous" },
{ provider: "bluesminds", modelId: "grok-3", displayName: "Grok-3 (VIP)", monthlyTokens: 7200000, creditTokens: 0, freeType: "recurring-daily", poolKey: "bluesminds", tos: "ambiguous" },
{ provider: "bluesminds", modelId: "qwen-max", displayName: "Qwen Max (VIP)", monthlyTokens: 7200000, creditTokens: 0, freeType: "recurring-daily", poolKey: "bluesminds", tos: "ambiguous" },
{ provider: "bytez", modelId: "meta-llama/Llama-3.3-70B-Instruct", displayName: "meta-llama/Llama-3.3-70B-Instruct", monthlyTokens: 0, creditTokens: 1000000, freeType: "recurring-credit", poolKey: "bytez", tos: "ambiguous" },
{ provider: "bytez", modelId: "mistralai/Mistral-7B-Instruct-v0.3", displayName: "mistralai/Mistral-7B-Instruct-v0.3", monthlyTokens: 0, creditTokens: 1000000, freeType: "recurring-credit", poolKey: "bytez", tos: "ambiguous" },
{ provider: "bytez", modelId: "Qwen/Qwen2.5-72B-Instruct", displayName: "Qwen/Qwen2.5-72B-Instruct", monthlyTokens: 0, creditTokens: 1000000, freeType: "recurring-credit", poolKey: "bytez", tos: "ambiguous" },
// evidence: public-page https://docs.bytez.com/model-api/docs/billing (2026-10-15).
// Free open models are limited to 7B; credits refresh every 4 weeks and bill per inference second.
// Token allowance is unpublished: both token counters are 0 (unknown), not a claim of zero credits.
// Larger 70B/72B tiers require purchased credits and are not cataloged as free.
{ provider: "bytez", modelId: "mistralai/Mistral-7B-Instruct-v0.3", displayName: "mistralai/Mistral-7B-Instruct-v0.3", monthlyTokens: 0, creditTokens: 0, freeType: "recurring-credit", poolKey: "bytez", tos: "ambiguous" },
// #11773: cerebras.ai/pricing (2026-09-03) is a one-time $5 signup credit
// gated on a payment method, 30-day expiry — not the old no-card 1M/day
// trial. creditTokens stays 0 because Cerebras publishes dollars, not a
@@ -303,8 +305,8 @@ export const FREE_MODEL_BUDGETS: FreeModelBudget[] = [
{ provider: "muse-spark-web", modelId: "muse-spark-contemplating", displayName: "Muse Spark Contemplating", monthlyTokens: 0, creditTokens: 0, freeType: "keyless", poolKey: "muse-spark-web", tos: "avoid" },
{ provider: "nebius", modelId: "meta-llama/Llama-3.3-70B-Instruct", displayName: "Llama 3.3 70B Instruct", monthlyTokens: 0, creditTokens: 1000000, freeType: "one-time-initial", poolKey: "nebius", tos: "caution" },
{ provider: "nlpcloud", modelId: "llama-3-8b-instruct", displayName: "Llama 3 8B", monthlyTokens: 0, creditTokens: 0, freeType: "recurring-monthly", poolKey: "nlpcloud", tos: "avoid" },
{ provider: "nous-research", modelId: "Hermes-4-405B", displayName: "Hermes 4 405B (Nous Research)", monthlyTokens: 0, creditTokens: 0, freeType: "recurring-credit", poolKey: "nous-research", tos: "ambiguous" },
{ provider: "nous-research", modelId: "Hermes-4-70B", displayName: "Hermes 4 70B (Nous Research)", monthlyTokens: 0, creditTokens: 0, freeType: "recurring-credit", poolKey: "nous-research", tos: "ambiguous" },
// evidence: public-page https://portal.nousresearch.com/ (2026-10-15): Free models only,
// but no named Hermes-4 coverage; omit those unverified rows until evidence establishes access.
{ provider: "novita", modelId: "ai-ai/llama-3.1-8b-instruct", displayName: "Llama 3.1 8B", monthlyTokens: 0, creditTokens: 500000, freeType: "one-time-initial", poolKey: "novita", tos: "caution" },
{ provider: "nscale", modelId: "moonshotai/Kimi-K2.5", displayName: "moonshotai/Kimi-K2.5", monthlyTokens: 0, creditTokens: 5000000, freeType: "one-time-initial", poolKey: "nscale", tos: "caution" },
{ provider: "nscale", modelId: "Qwen/Qwen3-235B-A22B-Instruct-2507", displayName: "Qwen/Qwen3-235B-A22B-Instruct-2507", monthlyTokens: 0, creditTokens: 5000000, freeType: "one-time-initial", poolKey: "nscale", tos: "caution" },

View File

@@ -0,0 +1,61 @@
import assert from "node:assert/strict";
import test from "node:test";
import {
FREE_MODEL_BUDGETS,
computeFreeModelTotals,
} from "../../open-sse/config/freeModelCatalog.ts";
import { isFreeModel, selectModelsForImport } from "../../src/shared/utils/freeModels.ts";
const bytezPaidModels = ["meta-llama/Llama-3.3-70B-Instruct", "Qwen/Qwen2.5-72B-Instruct"];
const nousUnverifiedModels = ["Hermes-4-405B", "Hermes-4-70B"];
const mistral7B = "mistralai/Mistral-7B-Instruct-v0.3";
for (const [provider, models] of [
["bytez", bytezPaidModels],
["nous-research", nousUnverifiedModels],
] as const) {
for (const id of models) {
test(`${provider}/${id} does not receive free status from unsupported catalog evidence`, () => {
assert.equal(isFreeModel(provider, { id }), false);
assert.equal(
isFreeModel(provider, { id, isFree: false, pricing: { prompt: 1, completion: 2 } }),
false
);
});
}
}
test("Bytez free-only import keeps the documented 7B tier and removes paid unlock tiers", () => {
const fetched = [...bytezPaidModels, mistral7B].map((id) => ({ id }));
const result = selectModelsForImport("bytez", fetched, true);
assert.deepEqual(result.models, [{ id: mistral7B }]);
assert.equal(result.freeFilterEmpty, false);
assert.deepEqual(selectModelsForImport("bytez", fetched, false).models, fetched);
});
test("unverified Hermes rows cannot populate a Nous free-only import", () => {
const fetched = nousUnverifiedModels.map((id) => ({ id }));
const result = selectModelsForImport("nous-research", fetched, true);
assert.deepEqual(result.models, []);
assert.equal(result.freeFilterEmpty, true);
assert.deepEqual(selectModelsForImport("nous-research", fetched, false).models, fetched);
});
test("Bytez recurring credits retain free access without an invented token allowance", () => {
const entries = FREE_MODEL_BUDGETS.filter((entry) => entry.provider === "bytez");
assert.equal(entries.length, 1);
assert.equal(entries[0].modelId, mistral7B);
assert.equal(entries[0].freeType, "recurring-credit");
assert.equal(entries[0].monthlyTokens, 0);
assert.equal(entries[0].creditTokens, 0);
assert.equal(isFreeModel("bytez", { id: mistral7B }), true);
const totals = computeFreeModelTotals({ entries });
assert.equal(totals.steadyWithRecurringCreditsTokens, 0);
assert.equal(totals.firstMonthRealisticTokens, 0);
});
test("verified free models on other providers keep their existing classification", () => {
const entry = FREE_MODEL_BUDGETS.find((row) => row.provider === "groq");
assert.ok(entry);
assert.equal(isFreeModel(entry.provider, { id: entry.modelId }), true);
});

View File

@@ -48,20 +48,8 @@ test("nous-research Hermes-4-405B is labelled 405B (Nous Research)", () => {
assert.equal(model.name, "Hermes 4 405B (Nous Research)");
});
test("nous-research free-model catalog display names match their modelId's parameter size", () => {
test("unverified Nous Hermes rows stay outside the free catalog", () => {
// #16781: the public Free plan does not identify these registry models as covered.
const rows = FREE_MODEL_BUDGETS.filter((model) => model.provider === "nous-research");
assert.ok(rows.length > 0, "FREE_MODEL_BUDGETS must list nous-research models");
for (const row of rows) {
const idSize = extractParamSize(row.modelId);
assert.ok(idSize, `nous-research free-catalog modelId "${row.modelId}" must encode a size`);
const nameSize = extractParamSize(row.displayName);
assert.equal(
nameSize,
idSize,
`nous-research free-catalog "${row.modelId}" displayName "${row.displayName}" advertises ` +
`${nameSize} but the modelId says ${idSize}`
);
}
assert.deepEqual(rows, []);
});