Claude Sonnet 5.5 vs Gemini 4 Argon
| Claude Sonnet 5.5 | Gemini 4 Argon | |
|---|---|---|
| Announced | ||
| Released | ||
| Input / 1M tokens | $2 | $2$4 from 1 Dec 2026 |
| Cached input / 1M tokens | $0.10 | $0.10$0.20 from 1 Dec 2026 |
| Output / 1M tokens | $10 | $10$20 from 1 Dec 2026 |
| Context window | 1M | 1M |
| Output limit | 128K | 65K |
| Benchmarks | ||
| Terminal-Bench 4.0 | 70.6%1 | 57.4% |
| Terminal-Bench Science 0.1 | 59.9%2 | 57.6% |
| AutomationBench 1.0.6 Private set | 44.7%3 | 51.3%4 |
| OSWorld 2.1 Offline set (82 tasks), partial score | 83.9%5 | 74.7%6 |
| FrontierSWE v2 | 61.9%7 | 55.0%8 |
| DeepSWE v1.1 | 71.0%9 | 77.9% |
| Chartography No tools | 61.6% | 71.6%10 |
| Claude Sonnet 5.5 | Gemini 4 Argon | |
|---|---|---|
| Other benchmarks | ||
| Humanity's Last Exam No tools, full set (2,500 questions) | 56.9% | No data |
| Humanity's Last Exam With tools, full set (2,500 questions) | 64.5% | No data |
| SWE-bench Pro Set not stated | 81.3%11 | No data |
| OSWorld 2.1 Full set (108 tasks), partial score | 80.1%12 | No data |
| OSWorld 2.1 Full set (108 tasks), strict pass rate | 43.5%13 | No data |
| OSWorld 2.1 Offline set (82 tasks), strict pass rate | 48.8%14 | No data |
| OSWorld 2.0 (v2026.08.08) Offline set, partial score | No data | 69.2%15 |
| FrontierCode 1.1 Main set | 52.1%16 | No data |
| FrontierCode 1.1 Extended set | 64.4%17 | No data |
| Chartography With tools | 90.2%18 | No data |
| SEC-bench Pro 260505 | No data | 83.4%19 |
| Agents' Last Exam v1 Pass rate, set not stated | No data | 39.5%20 |
| Agents' Last Exam v1.1 Score, all tasks (122) | No data | 69.0%21 |
| GraphWalks BFS, up to 128K | No data | 99.7%22 |
| GraphWalks BFS, 256K + 1M problems | No data | 84.2% |
Notes
- Terminal-Bench 4.0, Claude Sonnet 5.5: adaptive thinking
- Terminal-Bench Science 0.1, Claude Sonnet 5.5: Claude Code in --bare mode · max effort · standard error ±4.7 points
- AutomationBench 1.0.6, Private set, Claude Sonnet 5.5: Run by Zapier · max effort · default fallbacks on
- AutomationBench 1.0.6, Private set, Gemini 4 Argon: Run by Zapier · set named in the evaluation methodology the post cites
- OSWorld 2.1, Offline set (82 tasks), partial score, Claude Sonnet 5.5: max effort · mean of 5 attempts
- OSWorld 2.1, Offline set (82 tasks), partial score, Gemini 4 Argon: Gemini CUA harness · highest thinking · max of 3 runs, one attempt per run
- FrontierSWE v2, Claude Sonnet 5.5: Run by Proximal · Proximal's agent harness · max reasoning effort · mean of 5 trials per task · 20-hour budget
- FrontierSWE v2, Gemini 4 Argon: Run by Proximal · Proximal's public leaderboard
- DeepSWE v1.1, Claude Sonnet 5.5: average of 5 trials
- Chartography, No tools, Gemini 4 Argon: Run by Surge AI · Surge AI's public leaderboard
- SWE-bench Pro, Set not stated, Claude Sonnet 5.5: max effort · adaptive thinking · average of 5 trials
- OSWorld 2.1, Full set (108 tasks), partial score, Claude Sonnet 5.5: max effort · mean of 5 runs · September 10, 2026 task files
- OSWorld 2.1, Full set (108 tasks), strict pass rate, Claude Sonnet 5.5: max effort · mean of 5 runs · September 10, 2026 task files
- OSWorld 2.1, Offline set (82 tasks), strict pass rate, Claude Sonnet 5.5: max effort · mean of 5 attempts
- OSWorld 2.0 (v2026.08.08), Offline set, partial score, Gemini 4 Argon: Gemini CUA harness · max of 3 runs, one attempt per run
- FrontierCode 1.1, Main set, Claude Sonnet 5.5: Run by Cognition · Claude Code harness · xhigh effort
- FrontierCode 1.1, Extended set, Claude Sonnet 5.5: Run by Cognition · Claude Code harness · xhigh effort
- Chartography, With tools, Claude Sonnet 5.5: max effort · adaptive thinking · average of 5 runs · container and image-cropping tool · Gemini 3.5 Flash grader
- SEC-bench Pro 260505, Gemini 4 Argon: Antigravity harness · pass@1, mean of 3 attempts · 183-task JavaScript engine split (May 2026) · LLM judge for crashes in both builds
- Agents' Last Exam v1, Pass rate, set not stated, Gemini 4 Argon: ALE-Claw harness · 5-hour window · safety filters on
- Agents' Last Exam v1.1, Score, all tasks (122), Gemini 4 Argon: ALE-Claw harness · 5-hour window · October 17, 2026 task files · safety filters on
- GraphWalks, BFS, up to 128K, Gemini 4 Argon: 650 problems up to 128K tokens, per the model evaluation document