AI engineering, October 30, 2026

Top News · 1,291 HN points

Mistral positions Large 4, also called Le Chonk, as its flagship open-weight model. In Surge Code's blind coding evaluations, it ranked behind only Claude Opus 5.5, ahead of GLM-5.3, Kimi K3, and DeepSeek V4.1 Flash among the evaluated models.

Here's what changed:

• Artificial Analysis' Coding Agent Index gives it 48%, tied with GLM-5.3.
• Artificial Analysis separately reports 47% fewer tokens per task on average.
• API pricing is $0.70 per million input tokens, $0.07 per million cached input tokens, and $2.10 per million output tokens.
• Terminal-Bench 4.0 reaches 41%, up 13 percentage points from preview.
• Low, Medium, and High reasoning modes let you trade performance against cost. High is the default.

One catch: The Mistral Ecosystem Open-Weight license allows use, modification, fine-tuning, and commercialization, but commercial Model-as-a-Service operators need a separate license.

Try it: Start in Mistral Studio or use the Mistral API. You can also download the weights from Hugging Face and self-host with vLLM or SGLang.

Sources: mistral.ai

Top News · Reported · tomshardware.com

The RTX 5080's reported online price floor has reached $1,999. The card launched at $999, and its current online floor matches the RTX 5090's launch MSRP.

Here's what current listings look like:

• Best Buy had a PNY RTX 5080 for $1,999, the cheapest in-stock card found at a reputable online seller.
• Micro Center listed a Gigabyte card at $1,979.99 for in-store pickup, depending on location.
• Newegg had an Asus card for at least $2,299.99.
• The RTX 5080 pairs Nvidia's GB203 GPU with 10,752 CUDA cores and 16GB of GDDR7 VRAM.

One catch: Stock is limited, and much of the online supply comes from third-party sellers.

Why care? If you're buying an RTX 5080, check in-stock prices rather than budgeting from its $999 launch MSRP.

Sources: tomshardware.com

Top News

Eligible ChatGPT Pro and Business subscriptions can now offset paid API Platform usage. OpenAI's new benefit gives eligible subscribers recurring API credits after they link an API organization.

Here's what you get:

• Pro plans receive $100, $200, or $500 per month, depending on tier.
• Business Standard gets $20 per paid seat and Business Premium gets $100 per paid seat, pooled up to $500 per workspace monthly.
• Credits cover Responses, Agents, Decisions, GPT-Live/Realtime, Images, and built-in tools.
• Grants arrive automatically at subscription renewal after the initial claim.

One catch: You must own the subscription or workspace and API organization, use the same email, and have a paid subscription at least seven days old. Trials and complimentary subscriptions do not qualify. The first grant is prorated, unused credits expire at cycle end, and Codex API-key usage is excluded.

Try it: In API Platform, select an organization you own, open Billing, and link your subscription or workspace.

Sources: developers.openai.com · developers.openai.com

1. Amazon SageMaker AI cuts GPU idle time during image pulls by up to 90%
2. Mastra’s open-source agent memory hits 80.6% on BEAM at 1M tokens
3. LiteRT in Google Play services runs GPU inference about 1.4x faster than its legacy delegate
4. Scale AI releases FineART, an open robotics dataset with 533,913 labeled steps
5. Impeccable checks interface designs with 62 deterministic rules and no API key · 99 Stars today
6. WeVisCut edits videos with nearly 33x fewer tokens per successful output minute than Claude Opus 5 · 13 Upvotes