GPT-5.5

Epoch said an AI-generated solution found a presentation for the absolute Galois group of the 2-adic numbers, marking the second FrontierMath open problem it says AI solved. The result was elicited with Fable 5 and also GPT-5.5 Pro, making it a benchmark milestone rather than a product release.

Moonshot launched Kimi K3 in Kimi products and API with 1M context, native multimodality, KDA/AttnRes, and weights promised by July 27. Benchmarks place it near frontier systems, but testers cite slow serving and usability caveats.

Morpheus gives models persistent simulation environments where rules, objectives, and consequences shift without resets. Early reports said frontier models leaned on pretraining heuristics.

Databricks published an internal coding-agent benchmark using tasks from its codebase. OpenAI, Anthropic, and GLM-5.2 models landed on its Pareto frontier, and the company argues teams should optimize cost per task rather than per token.

Posts say Sol, Terra, and Luna are set for a July 9 launch. One report says Sol was added to the Codex codebase as OpenAI’s strongest model for code, research, and documents.

ClaudeDevs reports Sonnet 5 with a Fable 5 advisor reached ~92% of Fable 5's SWE-bench Pro score at ~63% of the price. Other builders route implementation to Sonnet, Codex, GPT-5.5, or GLM workers.

CAIS and Scale’s Remote Labor Index reportedly put Fable 5 at 16.1% human-accepted freelance tasks, versus 8.3% for Opus 4.8 and 6.3% for GPT-5.5. The same report says AI judges overrated newer models, especially GPT-5.5 and Opus 4.8.

Fable users reported cost escalations from $300/day estimates to a single high-effort prompt above $130 and quota-drain complaints on Reddit. Users also reported automatic Opus fallback and safety refusals tied to bio/cyber safeguards.

Codex app code now references GPT-5.6 Sol, Terra, and Luna, while posts claim Sol Ultra reaches 91.9% on TerminalBench at lower cost. Treat release timing, limits, and benchmark claims as unofficial until OpenAI publishes details.

Vals and Artificial Analysis published independent Sonnet 5 results a day after launch, placing it just behind Opus 4.8 and Fable 5 while using far more turns than Sonnet 4.6. Lower token pricing did not make agentic tasks cheaper, and some finance benchmarks still triggered refusals.

Independent results put GLM-5.2 at the top of the open-model DeepSWE board and near the top on debate and post-train evals. Watch token use and long reasoning traces, which can offset its headline price advantage.

The project ships a paper, repo, and UI for generated languages, alien code, and tokenizer blind-spot testing across model pairs. Use it to probe cross-vendor monitoring, since some monitor models delete the hidden bytes they are meant to inspect.

TryCua and Snorkel opened Cua-Bench, a computer-use benchmark with 25 expert-authored KiCad tasks graded by exact netlist matches. The early results show frontier models still struggle with GUI execution, wiring completion, and self-checking, so treat benchmark wins as incomplete for real computer-use work.

Two days after Fable 5 went offline, developers started testing GLM-5.2, GPT-5.5, and multi-model panels against the kinds of one-shot frontend and greenfield builds Fable handled well. The early pattern is that replacements cover much of the work, but Fable still leads on UI taste and first-pass product completion.

Cognition introduced FrontierCode, a coding benchmark that grades mergeability and review quality instead of only unit-test passes, and the top model scored 13%. The result matters because it differs from SWE-Bench-style pass rates, and outside researchers are already questioning score variance and reproducibility.

Arena shipped Agent Mode, a benchmark that lets models use web search, bash, file writing, image generation, and follow-up questions, then ranks them on five live-session signals. It matters because agent evals move from static task sets to real user workflows, with GPT-5.5 High currently leading the leaderboard.

Independent users compared GPT-5.5/Codex with Opus 4.8/Claude Code using DeepSWE cost charts, GBA Eval runs, and long coding sessions. The split matters because engineers choosing a daily coding stack now have external quality-versus-cost evidence instead of only vendor launch claims.

Three days after Opus 4.8 launched, new tests and field reports added failed tool calls, Bash-specific breakdowns, and higher token burn to the complaint list. Users report materially worse cost and stability in long coding sessions, while DeepSWE and GBA Eval point in different directions.

Two days after launch, users and benchmarks pointed to write failures, sycophancy, lower security recall, and a 58% DeepSWE result. GPT-5.5 still leads on cost, output tokens, and pass@1 in shared coding-agent tests, so compare both before switching.

Anthropic released Claude Opus 4.8 across Claude, the API, and major clouds with higher coding scores and a cheaper 2.5x-speed Fast mode. Use it for coding workloads that want better benchmark performance without a price increase over 4.7.

OpenAI rolled a new GPT-5.5 Instant into ChatGPT and the API with less bullet-heavy output, better pacing, and higher multilingual quality. The update also replaces Canvas in GPT-5.5 Instant and Thinking with in-chat writing and code blocks, so users should migrate workflows while legacy models still keep Canvas temporarily.

DeepSWE launched a coding benchmark built from 113 original tasks across 91 repos and five languages, with GPT-5.5 leading at 70%. The setup is meant to better reflect repo search, multi-file edits, and verification in real agent workflows.

OpenAI said ChatGPT-linked Codex will drop GPT-5.2 and GPT-5.3-Codex on June 2, with GPT-5.5 becoming the default frontier model for free users. The API versions stay available, but the in-product model surface is being reduced for compute-fleet management.

Days after `/goal` workflows first surfaced, users showed the command also works in the Codex app and shared runs for SSH setup, mech-interp scripts, and recurring work that lasted hours or days. The evidence points to Codex being used as a long-running research and ops agent, though the app still lacks explicit `/goal` UI.

OpenAI said Codex’s GPT-5.5 degradation over the prior 48 hours came from two issues and it will reset usage limits after the fix. Users had reported looping runs, higher cache burn, and unstable sessions in active coding workflows.

OpenAI launched Daybreak, combining GPT-5.5, Codex workflows, repo scanning, threat modeling, and patch generation for cyber-defense teams. It packages frontier models into a continuous secure-software workflow, so teams can test whether it fits their response pipeline.

Engineers shared fresh measurements on GPT-5.5 cache reuse, /fast pricing, and bug-finding budgets after comparison posts for GPT-5.5 and Opus 4.7 led the coding round-up. The reports suggest Codex cost and quality now swing on cache behavior and effort settings as much as on list prices.

User posts and HN threads compared GPT-5.5 and Opus 4.7 across plan mode, frontend work, and 120K-context sessions. The split results mean token burn and instruction discipline matter as much as raw benchmark scores.

OpenAI introduced GPT-5.5-Cyber in limited preview for defensive security teams and paired it with GPT-5.5 plus Trusted Access for Cyber. The release matters because OpenAI is separating cyber-specific access and permissiveness from general-model access rather than treating security work as a normal prompting mode.

OpenAI is rolling GPT-5.5 Instant into ChatGPT as the default model and exposing it as gpt-5.5-chat-latest, alongside Memory Sources for personalized replies. The model also claims 52.5% fewer high-stakes hallucinations, so watch for behavior changes in production prompts.

Ahead of GitHub Copilot's June 1 usage-based billing switch, users documented GPT-5.5 sessions hitting 60M tokens and $221 across 15 messages on the legacy per-message plan. The examples show why flat message buckets break once single requests can run for hours and consume extreme token counts.

Developers posted side-by-side reports of faster one-shot fixes, 1.7B-token workdays, and fewer limit warnings with GPT-5.5 fast mode after OpenAI added Claude Code import. The comparisons matter because they turn migration talk into a concrete workflow choice.

ARC Prize published frontier-model results on ARC-AGI-3 and said GPT-5.5 and Opus 4.7 both stayed below 1%, with failures in world modeling, abstraction, and reward reinforcement. That shows strong coding and benchmark models still break on novel interactive reasoning tasks, and follow-up comparisons even had Opus 4.6 slightly ahead of 4.7.

ValsAI found that undocumented `tool_choice` behavior was skewing Terminal Bench 2 scores when no native tools were used, then reran the evals. The correction lifted GPT-5.5 by 11% to the top slot and showed how much harness settings can move coding-agent results.

Multiple summaries of the UK AISI report say GPT-5.5 roughly matches Claude Mythos Preview on long-horizon cyber tasks, including 2 of 10 end-to-end TLO completions. That matters because the model is broadly usable today, shifting cyber-workflow choices toward availability and mitigations rather than gated access alone.

OpenAI expanded Codex with role-based work-flows, app connections, in-app previews, and the `/goal` command, while also improving browser use by about 20%. The update lets Codex keep working across docs, slides, spreadsheets, and web actions instead of staying in a single coding thread.

Codex gained background macOS control, page inspection, image generation, plugins, artifacts, and follow-up automations. That gives it one agent thread for desktop apps, frontend debugging, and recurring work.

New evals and day-three user tests show GPT-5.5 performing well at low or medium reasoning, with benchmark gains over GPT-5.4 in coding-heavy use. That matters because stronger results no longer require xhigh runs, though some users still flag sycophancy.

Independent tools and platforms shipped GPT-5.5 support within a day of the API rollout, spanning IDEs, hosted research agents, enterprise stacks, and coding agents. That shortens evaluation time because teams can test the model inside existing workflows instead of rebuilding around a single OpenAI surface.

Users and third-party evals reported shorter runs, stronger long-context scores, and faster rollout into Cursor and other tools a day after GPT-5.5 hit the API. Higher per-token pricing may be partly offset by lower loop time and fewer tool-call stalls, so watch early bench data before changing defaults.

A day after GPT-5.5 and the new Codex workflows launched, developers reported one-shot bug fixes, longer unattended runs, and lower token use in real coding tasks. The early hands-on comparisons matter because they are already shifting some teams' default agent workflow away from Claude Code.

Cursor 3.2 added /multitask async subagents, improved worktrees, and multi-root workspaces, then paired the release with GPT-5.5 rollout at 72.8% on CursorBench. The update makes background agent orchestration a first-class IDE workflow instead of a blocking queue.

OpenAI added GPT-5.5 and GPT-5.5 Pro to the API and Playground with 1M context and Responses support. Partners including OpenRouter, Perplexity, GitHub Copilot, Vercel, Warp, and Devin rolled it out the same day, widening access beyond Codex.

OpenAI rolled out GPT-5.5 and GPT-5.5 Pro in ChatGPT and Codex, with higher scores on terminal, OS, cyber, and math evals than GPT-5.4. Codex also gained browser, document, and computer-use features for longer agent workflows.