Claude Opus 4.8vsgpt-oss-120b
Claude Opus 4.8 | gpt-oss-120b | |
|---|---|---|
| Specifications | ||
ParametersA rough measure of how big the model is. More parameters usually means more capable and more expensive to run, though it is a poor guide on its own — a smaller, newer model often beats a larger, older one. | — | 117B |
Context windowHow much text the model can hold in mind at once — your question, any documents you attach, the conversation so far, and its own reply. Go past it and the earliest part falls out of view. | 1M | 128k |
| API pricingUSD per 1M tokens · lower wins | ||
Input priceWhat you pay for everything you send the model — your question, plus any documents or earlier conversation you include with it. | $5.00 | — |
Output priceWhat you pay for the text the model writes back. It is normally the dearer half: producing an answer costs more than reading one. | $25.00 | — |
Cached input priceA reduced rate for text you send over and over. If every request starts with the same instructions or the same document, the provider keeps a copy ready and charges less to read it again. | $0.50 | — |
Cheapest inputLowest input rate across third-party providers, excluding the lab itself. The cheapest endpoint may run a quantised build or a shorter context — see "Available from" on the model page. | $5.00Amazon Bedrock | $0.03AkashML |
Cheapest outputLowest output rate across third-party providers, excluding the lab itself. May come from a different provider than the cheapest input. | $25.00Amazon Bedrock | $0.17AkashML |
| Benchmarks | ||
BullshitBench v2Nonsense detection — Given a confidently-worded but nonsensical prompt, does the AI spot that it makes no sense and push back — instead of playing along and inventing an answer? The score is how often it clearly called out the nonsense. Higher is better. | 95% | 11% |
SWE-Bench VerifiedCoding — Real coding tasks pulled from open-source projects — the AI has to find and fix actual bugs. A human-checked version of the original SWE-Bench. Higher is better. | 88.6% | 62.4% |
Humanity's Last Exam · no toolsMultidisciplinary reasoning — Humanity's Last Exam — extremely hard expert questions across many subjects, written so you can't just look up the answer. “No tools” means the AI answers on its own. Higher is better. | 49.8% | 14.9% |
Humanity's Last Exam · with toolsMultidisciplinary reasoning — Humanity's Last Exam — extremely hard expert questions across many subjects. “With tools” means the AI is allowed to search the web or run code while answering. Higher is better. | 57.9% | 19% |
| BenchmarksPublished by one model only | ||
Gray Swan IPI · k = 1Prompt injection robustness — Attackers hide malicious instructions inside content the AI reads — a web page, an email, a document — and try to hijack what it does. Gray Swan's indirect prompt injection benchmark measures how often such an attack succeeds when the attacker gets a single try. Lower is better. | 0.5% | — |
Gray Swan IPI · k = 10Prompt injection robustness — Attackers hide malicious instructions inside content the AI reads — a web page, an email, a document — and try to hijack what it does. This variant gives the attacker 10 tries and counts an attack as successful if any of them works. Lower is better. | 4.1% | — |
Gray Swan IPI · k = 15Prompt injection robustness — Attackers hide malicious instructions inside content the AI reads — a web page, an email, a document — and try to hijack what it does. This variant gives the attacker 15 tries and counts an attack as successful if any of them works. Lower is better. | 5.5% | — |
SWE-Bench ProAgentic coding — Can the AI fix real bugs in real software? It's handed actual problems from open-source projects and has to write code that genuinely solves them. Higher is better. | 69.2% | — |
SWE-Bench MultilingualMultilingual coding — Like SWE-Bench, but the coding problems span many programming languages, not just one. Tests how broadly the AI can code. Higher is better. | 84.4% | — |
DeepSWE 1.0Agentic coding — Artificial Analysis' independent test of deep, agentic software-engineering work — the AI has to plan and carry out substantial coding tasks end to end. Higher is better. | 55.8% | — |
Next.js EvalsNext.js coding — Vercel's open eval of how well AI coding agents build and migrate real Next.js apps — measured as the share of tasks the agent completes successfully. Higher is better. | 81% | — |
Frontier-Bench v0.1Agentic computer work — A hard, ever-evolving set of real computer tasks — coding, system administration, data work, and more — that an AI agent has to complete on its own. Run by the Harbor / Laude Institute team as the successor to Terminal-Bench (v0.1 is the first release of the task set). The score is the share of tasks solved. Higher is better. | 21.1% | — |
Terminal-Bench 4.0Agentic terminal coding — Can the AI work in a command-line terminal — running commands and finishing technical setup tasks the way a developer would? Version 4.0 recalibrated how much time, CPU and memory each task gets, removed eight tasks and fixed nineteen, so fewer runs fail for reasons that have nothing to do with the model. Scores are not comparable with earlier versions. Higher is better. | 23.64% | — |
Terminal-Bench 2.1Agentic terminal coding — Can the AI work in a command-line terminal — running commands and finishing technical setup tasks the way a developer would? Higher is better. | 74.6% | — |
BU BenchBrowser agent — Can the AI drive a real web browser to finish tasks — clicking, filling forms, and navigating sites the way a person would? Run by Browser Use on their BU Bench task set. Higher is better. | 74% | — |
BrowseCompWeb browsing — Can the AI browse the web and track down hard-to-find answers? Higher is better. | 84.3% | — |
GPQA DiamondScience — Graduate-level science questions in biology, physics, and chemistry — hard enough that subject-matter PhDs score around 65%. Higher is better. | — | 80.1% |
MMLUGeneral knowledge — A 57-subject multiple-choice exam — history, law, medicine, maths — that was the standard measure of how much a model knows from 2020 until roughly 2024, when frontier scores crowded into the high 80s and labs moved on to harder tests. The scores here were published years apart under different testing setups, so read them as a historical record rather than a like-for-like ranking. Higher is better. | — | 90% |
OSWorld-VerifiedAgentic computer use — Can the AI actually operate a computer — clicking, typing, and using real apps — to finish tasks on its own? Higher is better. | 83.4% | — |
Finance Agent v2Agentic financial analysis — Tests the AI on real financial-analysis work, like digging through reports and making sound decisions. Higher is better. | 53.9% | — |
Harvey's Legal Agent BenchmarkAgentic legal work — Harvey's test of whether an AI agent can complete real legal work — drafting and reviewing documents, working with spreadsheets and presentations, and navigating files the way a lawyer's assistant would. Higher is better. | 9.58% | — |
TaxEval v2Tax questions — A set of real tax questions created by Vals AI — can the AI give accurate answers about tax rules and filings? Higher is better. | 75.63% | — |
MedScribeMedical admin work — Can the AI support doctors with their administrative work, like notes and paperwork? Created by Vals AI. Higher is better. | 85.75% | — |
GDPval-AAKnowledge work — Measures how well the AI does economically valuable knowledge work, judged against human experts. Shown as a rating (like a chess Elo) — higher is better. | 1890 | — |
GDPval-AA v2Knowledge work — economically valuable knowledge work (v2, re-based Elo) | 1600 | — |
| Overview | ||
| Company | Anthropic | OpenAI |
| Release date | May 28 2026 | Aug 5 2025 |
| Access | Proprietary | Open Weight |
Other comparisons
Claude Opus 4.8vsGemini 3.7 Flashgpt-oss-120bvsGemini 3.7 FlashClaude Opus 4.8vsMuse Glimmergpt-oss-120bvsMuse GlimmerClaude Opus 4.8vsGrok 4.6gpt-oss-120bvsGrok 4.6Claude Opus 4.8vsDeepSeek-V4-Pro-0813gpt-oss-120bvsDeepSeek-V4-Pro-0813Claude Opus 4.8vsMistral Medium 3.5gpt-oss-120bvsMistral Medium 3.5Claude Opus 4.8vsKimi K3gpt-oss-120bvsKimi K3
Frequently asked questions
Claude Opus 4.8 leads gpt-oss-120b on 4 of the 4 benchmarks they both report (BullshitBench v2, SWE-Bench Verified, Humanity's Last Exam). Only Claude Opus 4.8 has a verified first-party API price: $5.00 per million input tokens and $25.00 per million output tokens. No pay-as-you-go API rate is tracked for gpt-oss-120b. gpt-oss-120b shipped 296 days before Claude Opus 4.8, so benchmark comparisons should account for the intervening progress.
Context windows are 1M (Claude Opus 4.8) vs 128k (gpt-oss-120b). Claude Opus 4.8 is proprietary, while gpt-oss-120b is open weight.
On BullshitBench v2, Claude Opus 4.8 leads at 95% vs gpt-oss-120b at 11%. On SWE-Bench Verified, Claude Opus 4.8 leads at 88.6% vs gpt-oss-120b at 62.4%. On Humanity's Last Exam · no tools, Claude Opus 4.8 leads at 49.8% vs gpt-oss-120b at 14.9%. On Humanity's Last Exam · with tools, Claude Opus 4.8 leads at 57.9% vs gpt-oss-120b at 19%.
Claude Opus 4.8 was released by Anthropic on May 28 2026.
gpt-oss-120b was released by OpenAI on Aug 5 2025.
Claude Opus 4.8 leads on SWE-Bench Verified — Claude Opus 4.8 88.6% vs gpt-oss-120b 62.4%.
Claude Opus 4.8 leads on Humanity's Last Exam · no tools — Claude Opus 4.8 49.8% vs gpt-oss-120b 14.9%.
Only Claude Opus 4.8 has a verified first-party API price: $5.00 per million input tokens and $25.00 per million output tokens. No pay-as-you-go API rate is tracked for gpt-oss-120b. Rates are pay-as-you-go API prices verified on August 18, 2026.
Claude Opus 4.8 has a 1M context window; gpt-oss-120b has 128k.
Claude Opus 4.8 is a proprietary model released by Anthropic. gpt-oss-120b is an open weight model released by OpenAI.