Inkling 2 Small vs Jais 3

Inkling 2 Small 20

Published rubric. No payment protocol (0). Per-million-token prices for serverless inference and training published without a login on 6 November (20). No free tier listed, and the quickstart asks for payment details before training, per the tinker listing (0). Sign-up is a browser flow (0). 0 + 20 + 0 + 0 = 20.

Jais 3 60

Self-hosted rule. Apache-2.0 weights with nothing to buy and no hosted option from the makers (20). Free to run with no card (20). Downloads are ungated and anonymous, so an agent can fetch the weights with no sign-up (20). No payment protocol (0). 20 + 20 + 20 + 0 = 60.

Inkling 2 Small 57

Platform readings are the tinker listing's. Revocable API keys from the console, with no per-key scopes found (20). Org, team and project permissions with no read-only role (5). The model returns its own completions, no third-party content (10). get_audit_log() for admins (15). security.txt with a reporting address and no stated acknowledgement time (7). For the model reading, Tinker's terms commit not to use Customer Content for training, recorded under transparency. 20 + 5 + 10 + 15 + 7 = 57.

Jais 3 31

Read as software the owner runs. No account or key, ungated downloads, and server authentication is the owner's to add (12 of 30). A model that calls whatever tools the operator wires in. The cards rule high-stakes decisions without human oversight out of scope, but nothing in the model or template enforces approval (6 of 20). The template renders tool results in their own tool role, but string results go in as raw text (only non-string content is serialised with tojson), no escaping of special-token sequences in caller or tool text was found, and no injection guidance or results are published (4 of 15). No call log of its own. Visibility depends on the serving stack (5 of 15). Every launch command passes --trust-remote-code, so modeling_k2_horizon.py from the repository runs on the host. We read it and found only torch and transformers imports. No security.txt on ifm.ai (404 on 17 October, not re-read), no security contact on the cards and no safety evaluation found. Since 25 October the cards link a technical report on ifm.ai, which answered a Cloudflare bot check on 4 November, so what it covers is unread (4 of 20). 12 + 6 + 4 + 5 + 4 = 31.

Inkling 2 Small 30

Model reading, for the surface it's called on, Tinker's serverless inference and OpenAI-compatible API. No status page found, per the tinker listing's checks of 1 October and 4 November (0), so no readable incident history (5). The only published limit is the in-flight sample cap of 2,000, per the tinker listing's 1 October check; the OpenAI-compatible page gives no numbers and says higher or more stable throughput needs the waitlist (10). The SDK retries sampling with stable request IDs, per the tinker listing's reading of the SDK repository; the OpenAI-compatible page has no retry guidance (15). No SLA found (0). Serverless inference is 'currently in beta' on the models and pricing page and the OpenAI-compatible API is labelled Beta, read 6 November (0). 0 + 5 + 10 + 15 + 0 + 0 = 30.

Jais 3 48

Read with the local-software lines, since the owner serves the weights. Ungated weights on Hugging Face with stated runtimes (vLLM 0.30.0 or later, SGLang 0.5.20 or later, Transformers validated at 5.15.0, mlx-lm with shipped model code), but every launch command trusts remote code. The GGUF card still says to build llama.cpp from master or wait for a release, though llama.cpp's own tags carry K2 Horizon support from b11454 (6 October 2026) (16 of 20, held). No CI or test suite of the makers' was found, and the code repository is 'coming soon'. SGLang's cookbook reports a validation pass of the tool-call parser on all six checkpoints, and IFM's MLX cards describe conversion checks against the PyTorch code (3 of 25). No crash reports. On 4 November the six main repositories held seven discussions, six of them closed citation updates and one open MoVA-36B-A4B thread from 26 October with no reply yet. Two template fixes for OpenAI-style message history landed on 14 October, before the launch (15 of 25). No changelog or version tags, only public commit messages (2 of 15). Released as named final checkpoints with no preview label, though mlx-lm v0.32.2 still has no K2 Horizon model code of its own (12 of 15). 16 + 3 + 15 + 2 + 12 = 48.

Inkling 2 Small 70

Model reading. No OpenAPI file of its own; the OpenAI-compatible chat and completions endpoints and the typed Python SDK are the contract, as the tinker listing reads it (15). llms.txt on the docs host, per the tinker listing's 1 October check (10). The model card states intended use (agentic and tool-use systems, coding assistants, chatbots, RAG), inputs, outputs and limitations, and the launch post its benchmarks (16 of 20). Model IDs, context lengths and the reasoning_effort values (none to xhigh, or 0.0 to 0.99) are documented; the tool-call format and structured output aren't, for this model (8 of 15). A Python example and an HTTP 400 for models without reasoning_effort, with no error page (6 of 15). A dated Tinker changelog and a model deprecations page, per the tinker listing (15). 15 + 10 + 16 + 8 + 6 + 15 = 70.

Jais 3 54

Read for a model the owner serves. No specification of its own. The contract is the Jinja chat template, which validates tool JSON Schemas (required fields must exist in properties) and rejects unknown options, plus the OpenAI-compatible servers in vLLM and SGLang with the k2_horizon parsers (10 of 25). The cards are Markdown on Hugging Face, and SGLang's docs have an llms.txt, but no llms.txt or Markdown docs from IFM were found and ifm.ai couldn't be read (5 of 10). Each card states intended use, four out-of-scope uses, limitations and per-size serving advice (15 of 20). reasoning_effort and tool_call_format are enums enforced with errors, and tools take JSON Schema. The 375B card describes low as no reasoning and omits medium, which the template accepts, and structured output isn't documented (10 of 15). Python examples for a reasoning call and a tool loop, commands for five runtimes, and named template errors, but no error reference for a served model (11 of 15). No changelog or version scheme. Training-stage branches and public commit history (3 of 15). 10 + 5 + 15 + 10 + 11 + 3 = 54.

Inkling 2 Small 75

Model reading. Released 5 November 2026 (30). Churn and notice in place of release counts: Inkling, Inkling-Small (30 July) and Inkling 2 Small in about four months, and a deprecations page with dated retirements and a promise to 'aim to give advance notice' by email, with no stated period, per the tinker listing (10 of 20). A public changelog and a support address and Discord, per the tinker listing (10 of 15). Current official SDKs, Tinker's Python SDK and the OpenAI-compatible endpoint (15). SDK and cookbook CI, per the tinker listing (10). 30 + 10 + 10 + 15 + 10 = 75.

Jais 3 62

Read for an open-weight model released on 15 October 2026. The last change to the template or code was on 15 October, within 30 days. The cards gained a technical report link on 25 October and a citation update on 27 October, which change no file a runtime loads (30). One family release, with dated builds before it (FP8 from 29 September, MLX from 1 October, GGUF on 12 and 13 October) and template fixes on 14 and 15 October, but no deprecation notices or successor history to judge. K2 Horizon came on 1 September and Jais 2 in December 2025 (10 of 20, the model reading of this line, held). One outside question, on MoVA-36B-A4B on 26 October, had no reply on 4 November, and IFM's GitHub repositories weren't read. The template fixes match the problem IFM's MLX cards describe (7 of 25, held). Recipes in vLLM's and SGLang's own docs, K2 Horizon support in llama.cpp's tagged builds from b11454 and FP8, GGUF and MLX builds from IFM, but no native mlx-lm support (11 of 15). Current dependencies (Transformers 5.13 to 5.15, vLLM 0.30), with no CI visible (4 of 10). 30 + 10 + 7 + 11 + 4 = 62.

Inkling 2 Small 64

Editorial half. Closed for now under the Tinker Terms of Use and the Model Acceptable Use Policy, per the model card, with Apache 2.0 weights promised the week after launch (15). Terms commit not to use Customer Content for training, keep it only as needed, with a DPA and deletion on request; the privacy notice wasn't re-read, per the tinker listing (15). Dated deprecation notices with no stated period (15). The DPA annex names Google Cloud, Cloudflare, WorkOS and OpenAI with service and location (15). 15 + 15 + 15 + 15 = 60.

Jais 3 55

Apache-2.0 in every card's metadata, with staged training checkpoints as branches and training logs linked on Weights & Biases (not opened). There is no LICENSE file and the code repository is still 'coming soon'. A technical report is linked from 25 October but sits behind a bot check, so its data description is unread (24 of 30). Inputs stay on the owner's hardware because nothing is hosted, but no statement says so and no privacy document governs the weights (15 of 30). No deprecation policy, dated notices or version scheme (0 of 20). Safetensors weights carry no code, and the remote modelling and configuration code we read imports only torch, transformers and huggingface_hub, with no network calls. No telemetry statement (12 of 20). 24 + 15 + 0 + 12 = 51.