Claude Opus 5.5 vs Gemini 4 Argon
Full profiles: Claude Opus 5.5 (Anthropic) and Gemini 4 Argon (Google): every benchmark, price and rank.
Updated · Data from 10 sources
As of 21 Oct 2026, Claude Opus 5.5 (max) and Gemini 4 Argon (high) are neck and neck on quality (72.7 vs 71.0 on the BenchLeader Index, inside the ±6.2 margin of error), and Gemini 4 Argon (high) costs 2.0× less, which makes it the better buy for most workloads. Gemini 4 Argon (high)'s current price runs until 30 Nov 2026; from 1 Dec 2026 both cost about $8.00 per 1M tokens blended, so then pick on use case: Claude Opus 5.5 for coding (82 vs 72), Gemini 4 Argon for agents & tools (72 vs 64).
Claude Opus 5.5 benchmarks max Anthropic
Gemini 4 Argon benchmarks high GoogleOpen in the comparison toolWhich to pick for…
The pick's numbers come first. Category scores on the BenchLeader scale (50 is the average model, 15 points one standard deviation). High confidence means a lead of half a standard deviation or more; under one point is called either.
| Use case | Pick | Why | Confidence |
|---|---|---|---|
| Coding | Claude Opus 5.5 | 82 vs 72 category score; Arena WebDev 1807 vs 1678 | High |
| Agents & tool use | Gemini 4 Argon | 72 vs 64 category score; Arena Agent 12.1 vs 20.4 | High |
| Maths | Claude Opus 5.5 | 74 vs 71 category score; FrontierMath Tiers 1–3 91.2% vs 86.0% | Medium |
| Reasoning | Either | 76 vs 75 category score; GPQA Diamond 90.6% vs 96.4% | Either |
| Knowledge | Claude Opus 5.5 | 79 vs 77 category score; AA-Omniscience 46.4 vs 42.4 | Medium |
| Long documents | Claude Opus 5.5 | 68 vs 65 category score; AA-LCR 84.7% vs 79.7% | Medium |
| Chat quality (human preference) | – | Only Gemini 4 Argon has an Arena rating (73 category score) | Not comparable |
| Lowest cost | Gemini 4 Argon (until 30 Nov 2026) | $4.00 vs $8.00 per 1M blended; agentic coding session $0.160 vs $0.320; from 1 Dec 2026: $8.00 vs $8.00 | – |
| Fastest first answer | Gemini 4 Argon | 608.7 s (Claude Opus 5.5 (max)) vs 48.5 s (Gemini 4 Argon (high)) to the first answer token | – |
| Long single outputs | Claude Opus 5.5 | 128,000 vs 65,536 output tokens per request by default | – |
| High volume on a budget | Neither flagship | A smaller model is usually the better fit: Cheapest good LLMs, Fastest LLMs | – |
Key differences
- Claude Opus 5.5 (max) and Gemini 4 Argon (high) are neck and neck on quality (72.7 vs 71.0, within the ±6.2 margin).
- Claude Opus 5.5 (max) is stronger in coding, composite, knowledge, long context, maths, multimodal, reasoning.
- Gemini 4 Argon (high) is stronger in agents & tools, human preference.
- Gemini 4 Argon (high) is 2.0× cheaper ($4.00 vs $8.00 per 1M blended) until 30 Nov 2026, then $8.00 vs $8.00 from 1 Dec 2026.
At a glance
| Spec | Claude Opus 5.5 | Gemini 4 Argon |
|---|---|---|
| Developer | Anthropic | |
| Released | 22 Sept 2026 | 30 Sept 2026 |
| Generally available | 22 Sept 2026 | 13 Oct 2026 |
| Context window (tokens) | 1,000,000 | 1,048,576 |
| Max output (tokens) | 128,000 (up to 300K on the Message Batches API (beta)) | 65,536 (up to 1,048,576 with long decode) |
| Input types | text, image | text, image, video, audio, PDF |
| Reasoning effort settings | low, medium, high, xhigh, max | low, medium, high |
| Compared at | Claude Opus 5.5 (max) | Gemini 4 Argon (high) |
| Open weights | No | No |
| Price in / out / cached input per 1M | $4.00 / $20.00 / $0.200 | $2.00 / $10.00 / $0.100 |
| Announced price change | None announced | $4.00 / $20.00 / $0.200 from 1 Dec 2026 |
| API model ID | claude-opus-5-5 | gemini-4-argon |
| Where to use it | Claude API, Amazon Bedrock, Google Cloud, Microsoft Foundry | Gemini API (AI Studio), Gemini Enterprise Agent Platform, Gemini app |
Limits, availability and announced prices from the vendors' own pages: docs.anthropic.com, ai.google.dev, ai.google.dev pricing (checked 20 Oct 2026). Everything else is from the day's data.
Category by category
Category scores on the BenchLeader scale: 50 is the average evaluated model, 15 points is one standard deviation. A gap under one point is called level.
| Category | Claude Opus 5.5 (max) | Gemini 4 Argon (high) | Leader |
|---|---|---|---|
| Reasoning | 75.8 | 74.9 | level |
| Coding | 82.2 | 72.2 | Claude Opus 5.5 |
| Agents & tools | 63.5 | 72.0 | Gemini 4 Argon |
| Maths | 74.4 | 70.5 | Claude Opus 5.5 |
| Knowledge | 78.7 | 76.5 | Claude Opus 5.5 |
| Human preference | – | 72.9 | not comparable |
| Multimodal | 71.0 | – | not comparable |
| Long context | 67.5 | 64.9 | Claude Opus 5.5 |
| Composite | 94.2 | 88.5 | Claude Opus 5.5 |
Where sources disagree
Across the 35 benchmarks both have, Claude Opus 5.5 leads 29, Gemini 4 Argon leads 5 and 1 is a tie. Independent sources split: Claude Opus 5.5 leads on AA Intelligence Index v4.3.2 and Vals Index; Gemini 4 Argon leads on Arena Text.
| Overall index | Claude Opus 5.5 (max) | Gemini 4 Argon (high) | Leader |
|---|---|---|---|
| BenchLeader Index | 72.7 | 71.0 | neck and neck |
| AA Intelligence Index v4.3.2 | 57.6 | 52.6 | Claude Opus 5.5 |
| Vals Index | 70.2† | 69.2 | Claude Opus 5.5 |
| Arena Text | 1508† | 1524 | Gemini 4 Argon |
Each publisher runs its own tasks, harnesses and effort settings, so their overall verdicts can differ; BenchLeader reconciles them on one scale. † measured at another effort setting of the same model (the publisher did not test the one compared here).
Price, speed and latency
List prices blended at three input tokens per output token. Per-task prices use the token counts on the methodology page; suite cost is what Artificial Analysis paid to run its whole Intelligence Index.
| Measure | Claude Opus 5.5 (max) | Gemini 4 Argon (high) |
|---|---|---|
| Blended price per 1M tokens, until 30 Nov 2026 | $8.00 | $4.00 |
| Input / output per 1M | $4.00 / $20.00 | $2.00 / $10.00 |
| Blended price per 1M from 1 Dec 2026 | $8.00 | $8.00 |
| Input / output per 1M from 1 Dec 2026 | $4.00 / $20.00 | $4.00 / $20.00 |
| Cached input per 1M | $0.200 | $0.100 |
| Price per task: chat reply | $0.0076 | $0.0038 |
| Price per task: summarize a 30-page report | $0.060 | $0.030 |
| Price per task: code edit | $0.054 | $0.027 |
| Price per task: agentic coding session | $0.320 | $0.160 |
| Price per task: structured extraction | $0.012 | $0.0060 |
| Cost to run a full benchmark suite | $5.98 | $1.99 |
| Output speed | 95 tok/s | 83 tok/s |
| Time to first answer | 608.7 s | 48.5 s |
| Time to a complete answer | 613.9 s | 54.6 s |
Per token is not per task: on Artificial Analysis' full suite Claude Opus 5.5 (max) cost $5.98 and Gemini 4 Argon (high) $1.99, because Claude Opus 5.5 (max) used about 1.5× as many tokens for the same work, which a per-token price hides.
Each model is compared at its best-ranked reasoning-effort setting. Claude Opus 5.5 (max) scores 72.7 and takes 609 s to its first answer token; Claude Opus 5.5 (xhigh) scores 70.1 in 114 s, Claude Opus 5.5 (high) scores 69.4 in 34 s. Gemini 4 Argon (high) scores 71.0 and takes 49 s to its first answer token; Gemini 4 Argon scores 63.3 in 4 s. Higher effort means more thinking before the answer: better scores, longer waits and more tokens billed. How reasoning effort changes scores and cost.
Every benchmark
112 benchmarks, grouped by category; each name links to the board and its leaders. † measured at another effort setting of the same model, shown for reference and not counted as a win.
| Metric | Publisher | Claude Opus 5.5 (max) | Gemini 4 Argon (high) |
|---|---|---|---|
| BenchLeader Index | BenchLeader | 72.7 | 71.0 |
| Agents & tools score | BenchLeader | 63.5 | 72.0 |
| Coding score | BenchLeader | 82.2 | 72.2 |
| Composite score | BenchLeader | 94.2 | 88.5 |
| Knowledge score | BenchLeader | 78.7 | 76.5 |
| Long context score | BenchLeader | 67.5 | 64.9 |
| Maths score | BenchLeader | 74.4 | 70.5 |
| Multimodal score | BenchLeader | 71.0 | – |
| Reasoning score | BenchLeader | 75.8 | 74.9 |
| Human preference score | BenchLeader | – | 72.9 |
| Blended price $/M | BenchLeader | $8.00 | $4.00 |
| Output speed | BenchLeader | 95 tok/s | 83 tok/s |
| Time to first answer | BenchLeader | 608.7 s | 48.5 s |
| Context window | BenchLeader | 1,000,000 tokens | 1,048,576 tokens |
| Reasoning | |||
| GPQA Diamond | Epoch AI Benchmarking Hub | 90.6% | 96.4% |
| Arena Hard Prompts | Arena (LMArena) | 1533† | 1549 |
| LiveBench 2.0 Reasoning | LiveBench | 92.2%† | 95.1%† |
| LiveBench 1.0 Reasoning | LiveBench | 92.2% | – |
| Humanity's Last Exam (AA) | Artificial Analysis | 61.4% | 57.1% |
| ARC-AGI-1 | ARC Prize | 97.5% | 98.5% |
| ARC-AGI-2 | ARC Prize | 91.7% | 91.3% |
| ARC-AGI-3 | ARC Prize | – | 97.3% |
| CritPt | Artificial Analysis | 31.7% | 27.1% |
| MysteryMechanism | Vals AI | 49.5%† | 45.5% |
| Chess Puzzles | Epoch AI Benchmarking Hub | 67.0% | 59.0% |
| Mystery Game Puzzles | Epoch AI Benchmarking Hub | 71.0% | 64.0% |
| LMCA | Epoch AI Benchmarking Hub | 68.2% | 63.7% |
| DTBench | Epoch AI Benchmarking Hub | 98.9% | 97.3% |
| Coding | |||
| SciCode | SciCode | 66.9% | 61.8% |
| FrontierCode | Cognition | 54.6%† | 47.5%† |
| Arena Coding | Arena (LMArena) | 1548† | 1559 |
| Arena WebDev | Arena (LMArena) | 1807 | 1678 |
| LiveBench 2.0 Coding | LiveBench | 96.3%† | 89.7%† |
| LiveBench 1.0 Coding | LiveBench | 89.3% | – |
| IOI | Vals AI | 95.1%† | 100.0% |
| Code Migration | Vals AI | 66.7%† | 68.2% |
| ProgramBench | Vals AI | 18.5%† | 2.5% |
| Vibe Code Bench 1-100 | Vals AI | 30.4%† | 21.2% |
| Vibe Code Bench v1.1 | Vals AI | 90.3%† | 91.9% |
Show all 112 benchmarks
| Metric | Publisher | Claude Opus 5.5 (max) | Gemini 4 Argon (high) |
|---|---|---|---|
| Coding | |||
| MirrorCode | Epoch AI Benchmarking Hub | 77.4% | – |
| CursorBench | Cursor | 57.8% | 52.9% |
| ALE-Bench | Epoch AI Benchmarking Hub | 2147.4† | 2100.1 |
| FrontierSWE | Epoch AI Benchmarking Hub | 62.3% | 55.0% |
| Terminal-Bench 4.0 (AA) | Artificial Analysis | 59.6% | 57.1% |
| Startrise Frontend | Startrise | 71.1%† | – |
| FrontierCode 1.1 (AA) | Artificial Analysis | 52.2% | 43.5% |
| Agents & tools | |||
| Terminal-Bench | Terminal-Bench | 64.8% | – |
| Remote Labor Index | Scale AI / CAIS | 21.3%† | – |
| APEX-Agents | Mercor | 73.5% | 82.4%† |
| Arena Agent | Arena (LMArena) | 20.4 | 12.1 |
| LiveBench Long Agentic | LiveBench | 85.0%† | 69.9%† |
| LiveBench Browser | LiveBench | 80.0%† | 92.0%† |
| LiveBench Agentic Coding | LiveBench | 71.7% | – |
| Terminal-Bench 2.1 (Vals) | Vals AI | 87.6%† | – |
| HiL-Bench | Scale AI SEAL | 59.0%† | 62.0%† |
| GDPval-AA v2.1 | Artificial Analysis | 68.3% | 56.2% |
| ITBench SRE (AA) | Artificial Analysis | 38.2% | – |
| Analyst Agent (AA) | Artificial Analysis | 56.3% | 48.8% |
| CUA-bench | Vals AI | 14.0% | 4.8% |
| CyberBench | Vals AI | 55.4%† | 77.9% |
| Finance Agent v2 | Vals AI | 58.6%† | 65.4% |
| Harvey's Legal Agent Benchmark | Vals AI | 9.2%† | 27.1% |
| Legal Research Bench | Vals AI | 50.5%† | 54.8% |
| SREBench | Vals AI | 33.6%† | 44.3% |
| Tax Agent Bench | Vals AI | 45.6%† | 44.6% |
| Terminal-Bench 4.0 (Vals) | Vals AI | 65.2%† | 57.6% |
| Terminal-Bench Science | Vals AI | 47.1%† | 44.3% |
| Time Horizon Index: KSP | Vals AI | 91.3%† | 52.5% |
| Arena Search | Arena (LMArena) | 1211† | – |
| Vending-Bench 2 | Andon Labs | 9235.3† | 13718.2† |
| AutomationBench | Artificial Analysis | 69.5% | 77.5% |
| GDP.pdf (AA) | Artificial Analysis | 26.2% | 21.8% |
| Harvey LAB-AA v1.1 | Harvey | 4.2% | 3.9% |
| AA-Briefcase v1.1 | Artificial Analysis | 1803 | 1486 |
| Terminal-Bench-Science 0.1 (AA) | Artificial Analysis | 59.0% | 52.9% |
| CWE-Bench (AA) | Artificial Analysis | 58.3% | – |
| DeepSec-Bench (AA) | Artificial Analysis | 26.9% | – |
| CyberGym end-to-end (AA) | Artificial Analysis | 0.8% | – |
| FACTS Search | 86.5%† | 84.0%† | |
| Maths | |||
| FrontierMath Tiers 1–3 | Epoch AI Benchmarking Hub | 91.2% | 86.0% |
| FrontierMath Tier 4 | Epoch AI Benchmarking Hub | 95.0% | 78.0% |
| OTIS Mock AIME | Epoch AI Benchmarking Hub | 100.0% | 100.0% |
| ProofBench | Vals AI | 100.0% | 99.0%† |
| LiveBench Mathematics | LiveBench | 97.1% | – |
| Arena Maths | Arena (LMArena) | 1502† | 1529 |
| Knowledge | |||
| SimpleQA Verified | Epoch AI Benchmarking Hub | 72.2% | 76.0% |
| LiveBench 2.0 Data Analysis | LiveBench | 80.3%† | 85.2%† |
| LiveBench Research | LiveBench | 85.0%† | 83.1%† |
| LiveBench 1.0 Data Analysis | LiveBench | 80.3% | – |
| AA-Omniscience | Artificial Analysis | 46.4 | 42.4 |
| LegalBench | Vals AI | – | 88.3% |
| PRBench Finance | Scale AI SEAL | – | 49.9%† |
| PRBench Legal | Scale AI SEAL | 47.6%† | 49.4%† |
| BioMysteryBench | Vals AI | 79.3%† | 76.3% |
| Excel Modeling Benchmark | Vals AI | 75.9%† | 75.2% |
| MedCode | Vals AI | 49.8%† | 58.8% |
| MedScribe | Vals AI | 91.4%† | 87.4% |
| Public Benefits Bench | Vals AI | 70.6%† | 69.8% |
| EBR-bench | Epoch AI Benchmarking Hub | 71.4% | 64.8% |
| GDP.pdf | Epoch AI Benchmarking Hub | 30.6% | 28.4% |
| AA-Omniscience: accuracy | Artificial Analysis | 66.2% | 49.9% |
| AA-Omniscience: non-hallucination | Artificial Analysis | 41.4% | 84.9% |
| Instruction following | |||
| LiveBench Language | LiveBench | 86.3% | – |
| LiveBench Instruction Following | LiveBench | 65.7% | – |
| Arena Instruction Following | Arena (LMArena) | 1518† | 1527 |
| Human preference | |||
| Arena Text | Arena (LMArena) | 1508† | 1524 |
| Arena Creative Writing | Arena (LMArena) | 1519† | 1516 |
| Arena Multi-turn | Arena (LMArena) | 1501† | 1552 |
| Arena Longer Queries | Arena (LMArena) | 1523† | 1543 |
| Multimodal | |||
| Arena Vision | Arena (LMArena) | 1321† | – |
| LiveBench Design | LiveBench | 86.9%† | 73.7%† |
| MMMU-Pro | Artificial Analysis | 87.7% | – |
| SAGE | Vals AI | 45.8%† | 53.6% |
| AudioMC | Scale AI SEAL | – | 58.4% |
| AudioMC (text output) | Scale AI SEAL | – | 58.4% |
| Blueprint-Bench 2 | Epoch AI Benchmarking Hub | 51.2%† | 54.5%† |
| AA-3D | Artificial Analysis | 1191 | 1154 |
| Long context | |||
| AA-LCR | Artificial Analysis | 84.7% | 79.7% |
| MLCR | Artificial Analysis | 66.7% | – |
| Composite | |||
| Epoch Capabilities Index | Epoch AI Benchmarking Hub | 168.8† | 166.4† |
| LiveBench 2.0 | LiveBench | 86.5%† | 84.1%† |
| LiveBench 1.0 | LiveBench | 83.2% | – |
| AA Intelligence Index v4.3.2 | Artificial Analysis | 57.6 | 52.6 |
| Vals Index | Vals AI | 70.2† | 69.2 |
| AA Cyber Index | Artificial Analysis | 28.7 | – |
| Safety & honesty | |||
| FORTRESS | Scale AI SEAL | 5.8% | – |
Data as of 2026-10-21. Best configuration of each model; every score links to its source on the model pages.
What changed
- Gemini 4 Argon: list price moves to $4.00 / $20.00 per 1M input / output tokens (announced).
- Claude Opus 5.5 (max): 2 score revisions.
- Gemini 4 Argon (high): 1 new result (Harvey LAB-AA v1.1); 2 score revisions.
- Claude Opus 5.5 (max): BenchLeader Index 71.9 → 72.6.
- Claude Opus 5.5 (max): 1 new result (FORTRESS).
- Gemini 4 Argon (high): 3 new results (LMCA, DTBench, ALE-Bench).
- Claude Opus 5.5 (max): BenchLeader Index 71.3 → 71.9.
- Claude Opus 5.5 (max): 1 score revision.
- Gemini 4 Argon (high): 5 new results (Analyst Agent (AA), FrontierCode 1.1 (AA), Terminal-Bench-Science 0.1 (AA)…); 1 score revision.
- Gemini 4 Argon (high): 6 new results (EBR-bench, Time Horizon Index: KSP, Vibe Code Bench 1-100…); 1 score revision.
Related comparisons
- Claude Opus 5.5 vs Muse Wonder Full
- Claude Opus 5.5 vs GPT-6 Astra
- Claude Opus 5.5 vs Claude Fable 5.1
- Gemini 4 Argon vs Muse Wonder Full
- Gemini 4 Argon vs GPT-6 Astra
- Gemini 4 Argon vs Claude Fable 5.1
- Claude Opus 5.5 vs Gemini 3.8 Flash
- Claude Opus 5.5 vs GPT-6.1 Sol
Rankings: Best LLM overall · Best LLM for coding · Best LLM for agents · Best cheap LLMs
Claude Opus 5.5 vs Gemini 4 Argon: questions
- Is Claude Opus 5.5 better than Gemini 4 Argon?
- Claude Opus 5.5 (max) and Gemini 4 Argon (high) are neck and neck on quality (72.7 vs 71.0, within the ±6.2 margin). The BenchLeader Index combines every independent quality benchmark; the two are within the margin of error as of 2026-10-21, so check the category scores for your use.
- Which is cheaper, Claude Opus 5.5 or Gemini 4 Argon, and will that change?
- Gemini 4 Argon is cheaper today: $4.00 against $8.00 per million tokens, blended at three input tokens per output token. Yes, that changes: Gemini 4 Argon's list price changes on 1 Dec 2026, and from then they cost about the same ($8.00 vs $8.00 blended).
- Is Claude Opus 5.5 better than Gemini 4 Argon for coding?
- Claude Opus 5.5 scores higher in coding (82 vs 72 on the category index, where 50 is average). On Arena WebDev, the top shared coding benchmark, it is 1807 for Claude Opus 5.5 against 1678 for Gemini 4 Argon.
- Is Claude Opus 5.5 better than Gemini 4 Argon for agents and tool use?
- Gemini 4 Argon scores higher in agentic tasks (72 vs 64 on the category index, where 50 is average). On Arena Agent, the top shared agents & tools benchmark, it is 20.4 for Claude Opus 5.5 against 12.1 for Gemini 4 Argon.
- Which is better for long documents?
- Claude Opus 5.5 scores higher on long-context benchmarks (68 vs 65). Both take about 1M tokens: Claude Opus 5.5 1,000,000, Gemini 4 Argon 1,048,576. Per request, Claude Opus 5.5 writes up to 128,000 output tokens and Gemini 4 Argon up to 65,536 by default (up to 1,048,576 with long decode).
- Which has lower latency?
- Gemini 4 Argon (high) answers first: 48.5 s against 608.7 s to the first answer token, as Artificial Analysis measures it. Reasoning effort drives most of this: a higher setting thinks longer before it answers.
- Which is faster, Claude Opus 5.5 or Gemini 4 Argon?
- Claude Opus 5.5 streams faster: 95 against 83 output tokens per second.
- Do independent benchmarks agree on Claude Opus 5.5 vs Gemini 4 Argon?
- Across the 35 benchmarks both have, Claude Opus 5.5 leads 29, Gemini 4 Argon leads 5 and 1 is a tie. Independent sources split: Claude Opus 5.5 leads on AA Intelligence Index v4.3.2 and Vals Index; Gemini 4 Argon leads on Arena Text.
- When was Gemini 4 Argon released?
- Gemini 4 Argon became generally available on 13 Oct 2026, after first appearing on 30 Sept 2026.