GPT-6.1 Sol

Coding tests of Sol released two days after Sonnet 5.5, noting the odd behaviours seen while building with it.

Official summary of the Intelligence Index result, with pricing and cache-read changes.

Follow-up head-to-head with the creator reporting a clear winner after repeated trials.

Coding experiments with Sol from a channel that benchmarks models on real projects.

Same prompt run five times at low, medium, high, xhigh and max to show what extra reasoning effort changes in a Three.js-style scene.

Challenge video asking Sol to build a third-person horror game, using the Higgsfield API and MCP.

Early impressions of Sol at DevDay with the near-Astra claim tested at a fraction of the price.

Sol gets one hour, a Remotion project and a $5 OpenRouter budget to produce a video.

Official Higgsfield demo giving both models the same 3D game prototype task, with a video of the output.

Single-prompt attempt to recreate a WoW-style game, compared with GPT-6 results.

Hands-on test of agentic coding and frontend design asking if Sol can replace Opus 5.5.

Measurements from 24 hours of testing showing Sol 6.1 much slower than GPT-6 Sol and GPT-5.6 Sol at xhigh.

Scored comparison of coding output where Sol 6.1 lands near Grok 4.7 behind Astra and Opus 5.5.

Experiment pairing GPT-6.1 Sol as planner with a local Qwen3.8-27B coder on an RTX 3090, with timings and costs.

GPT-5.6, GPT-6 and GPT-6.1 Sol each build three games for a version-over-version comparison.

Port of Chromium Legacy 124 to Mac OS X Snow Leopard, done with Sol 6.1 and Astra.

Louis-Francois Bouchard's internal benchmark for GPT-6.1 Sol across effort levels with Elo, cost and time per script.

.NET 10 demo of GPT-6.1 Sol in Microsoft Foundry using Microsoft.Extensions.AI and the Responses API, covering PR triage, an approval-gated tool call and an evaluation harness.

JavaScript motion-graphics engine plus a skill that lets GPT models like Astra and Sol author polished programmatic videos.

Codex skill that lets Sol lead a task and selectively delegate pieces to Sol or Astra at high or xhigh reasoning effort.

Minecraft Java 1.20.1 magic mod written by GPT-6.1 Sol, shared as a worked example of model-authored game modding.

Two-mode Codex extension: Astra with adaptive reasoning, or Sol 6.1 adaptive and Luna max agents routed by Jev.

Short proofs for six Graffiti conjectures (198, 219, 252, 254, 584 and 712) written with Opus 5.5 and GPT-6.1 Sol.