{"slug": "13-models-and-4-agents-on-swe-tasks-go-java-python-rust-ts", "title": "13 Models and 4 Agents on SWE Tasks: Go, Java, Python, Rust, TS", "summary": "Fable 5 [high] Model led 13 models and 4 agents on SWE tasks with a 64.5% resolved rate, $4.40 cost per problem, and 2,518,308 tokens per problem, according to a benchmark table covering Go, Java, Python, Rust, and TypeScript. Grok 4.5 [high] Model followed at 63.8% with $1.47 per problem, while GPT-5.6 Sol [medium] Model achieved 62.3% at $0.85 per problem, and Junie Agent topped agents at 61.8% with $0.81 per problem.", "body_md": "# | Model | Resolved Rate (%) | Pass@5 (%) | Cost per Problem ($) | Tokens per Problem |\n|---|---|---|---|---|---|\n| 1 | Fable 5 [high] Model | 64.5% | 78.4% | $4.40 | 2,518,308 |\n| 2 | Grok 4.5 [high] Model | 63.8% | 77.5% | $1.47 | 2,429,424 |\n| 3 | Opus 5 [high] Model | 63.4% | 74.8% | $3.47 | 4,322,143 |\n| 4 | GLM-5.2 [high] Model | 62.9% | 81.1% | $1.40 | 5,524,892 |\n| 5 | GPT-5.6 Sol [medium] Model | 62.3% | 79.3% | $0.85 | 605,340 |\n| 6 | Junie Agent | 61.8% | 73.9% | $0.81 | 1,684,501 |\n| 7 | Claude Code Agent | 60.4% | 75.7% | $3.39 | 3,341,581 |\n| 8 | Codex Agent | 58.0% | 73.0% | $1.59 | 2,070,976 |\n| 9 | Sonnet 5 [high] Model | 56.8% | 74.8% | $1.43 | 4,645,617 |\n| 10 | Cursor Agent | 51.7% | 65.8% | $0.41 | 1,827,002 |\n| 11 | MiniMax M3 Model | 47.2% | 69.4% | $0.95 | 13,869,459 |\n| 12 | MiMo V2.5 Pro Model | 46.5% | 65.8% | $0.10 | 4,687,987 |\n| 13 | GPT-5.6 Luna [medium] Model | 43.6% | 59.5% | $0.11 | 395,522 |\n| 14 | DeepSeek-V4 Pro [high] Model | 40.2% | 64.0% | $0.15 | 3,955,414 |\n| 15 | Qwen3.6-27B Model | 31.2% | 57.7% | $0.62 | 3,078,970 |\n| 16 | Qwen3.6-35B-A3B Model | 24.7% | 43.2% | $0.27 | 3,726,061 |\n| 17 | Qwen3.5-35B-A3B Model | 17.1% | 36.9% | $0.99 | 5,939,227 |\n| 18 | Claude Opus 4.1 Model | N/A | N/A | N/A | N/A |\n| 19 | Claude Opus 4.5 Model | N/A | N/A | N/A | N/A |\n| 20 | Claude Opus 4.6-high Model | N/A | N/A | N/A | N/A |\n| 21 | Claude Opus 4.7-high Model | N/A | N/A | N/A | N/A |\n| 22 | Claude Opus 4.8-xhigh Model | N/A | N/A | N/A | N/A |\n| 23 | Claude Sonnet 3.5 Model | N/A | N/A | N/A | N/A |\n| 24 | Claude Sonnet 4 Model | N/A | N/A | N/A | N/A |\n| 25 | Claude Sonnet 4.5 Model | N/A | N/A | N/A | N/A |\n| 26 | Claude Sonnet 4.6 Model | N/A | N/A | N/A | N/A |\n| 27 | DeepSeek-R1-0528 Model | N/A | N/A | N/A | N/A |\n| 28 | DeepSeek-V3 Model | N/A | N/A | N/A | N/A |\n| 29 | DeepSeek-V3-0324 Model | N/A | N/A | N/A | N/A |\n| 30 | DeepSeek-V3-0324 Model | N/A | N/A | N/A | N/A |\n| 31 | DeepSeek-V3.1 Model | N/A | N/A | N/A | N/A |\n| 32 | DeepSeek-V3.2 Model | N/A | N/A | N/A | N/A |\n| 33 | DeepSeek-V4 Flash [high] Model | N/A | N/A | N/A | N/A |\n| 34 | Devstral-2-123B-Instruct-2512 Model | N/A | N/A | N/A | N/A |\n| 35 | Devstral-Small-2-24B-Instruct-2512 Model | N/A | N/A | N/A | N/A |\n| 36 | Devstral-Small-2505 Model | N/A | N/A | N/A | N/A |\n| 37 | Gemini 3 Flash Preview Model | N/A | N/A | N/A | N/A |\n| 38 | Gemini 3 Pro Preview Model | N/A | N/A | N/A | N/A |\n| 39 | Gemini 3.1 Pro Preview Model | N/A | N/A | N/A | N/A |\n| 40 | Gemini 3.5 Flash Model | N/A | N/A | N/A | N/A |\n| 41 | gemini-2.0-flash Model | N/A | N/A | N/A | N/A |\n| 42 | gemini-2.0-flash Model | N/A | N/A | N/A | N/A |\n| 43 | gemini-2.5-flash Model | N/A | N/A | N/A | N/A |\n| 44 | gemini-2.5-flash-preview-05-20 no-thinking Model | N/A | N/A | N/A | N/A |\n| 45 | gemini-2.5-flash-preview-05-20 no-thinking Model | N/A | N/A | N/A | N/A |\n| 46 | gemini-2.5-pro Model | N/A | N/A | N/A | N/A |\n| 47 | Gemma 4 31B Model | N/A | N/A | N/A | N/A |\n| 48 | gemma-3-27b-it Model | N/A | N/A | N/A | N/A |\n| 49 | GLM-4.5 Model | N/A | N/A | N/A | N/A |\n| 50 | GLM-4.5 Air Model | N/A | N/A | N/A | N/A |\n| 51 | GLM-4.6 Model | N/A | N/A | N/A | N/A |\n| 52 | GLM-4.7 Model | N/A | N/A | N/A | N/A |\n| 53 | GLM-4.7 Flash Model | N/A | N/A | N/A | N/A |\n| 54 | GLM-5 Model | N/A | N/A | N/A | N/A |\n| 55 | GLM-5.1 Model | N/A | N/A | N/A | N/A |\n| 56 | GLM-5.1 Model | N/A | N/A | N/A | N/A |\n| 57 | gpt-4.1-2025-04-14 Model | N/A | N/A | N/A | N/A |\n| 58 | gpt-4.1-2025-04-14 Model | N/A | N/A | N/A | N/A |\n| 59 | gpt-4.1-mini-2025-04-14 Model | N/A | N/A | N/A | N/A |\n| 60 | gpt-4.1-mini-2025-04-14 Model | N/A | N/A | N/A | N/A |\n| 61 | gpt-4.1-nano-2025-04-14 Model | N/A | N/A | N/A | N/A |\n| 62 | gpt-5-2025-08-07-high Model | N/A | N/A | N/A | N/A |\n| 63 | gpt-5-2025-08-07-medium Model | N/A | N/A | N/A | N/A |\n| 64 | gpt-5-2025-08-07-minimal Model | N/A | N/A | N/A | N/A |\n| 65 | gpt-5-codex Model | N/A | N/A | N/A | N/A |\n| 66 | gpt-5-mini-2025-08-07-high Model | N/A | N/A | N/A | N/A |\n| 67 | gpt-5-mini-2025-08-07-medium Model | N/A | N/A | N/A | N/A |\n| 68 | gpt-5.1-codex Model | N/A | N/A | N/A | N/A |\n| 69 | gpt-5.1-codex-max Model | N/A | N/A | N/A | N/A |\n| 70 | gpt-5.2-2025-12-11-medium Model | N/A | N/A | N/A | N/A |\n| 71 | gpt-5.2-2025-12-11-xhigh Model | N/A | N/A | N/A | N/A |\n| 72 | gpt-5.2-codex Model | N/A | N/A | N/A | N/A |\n| 73 | gpt-5.3-codex Model | N/A | N/A | N/A | N/A |\n| 74 | gpt-5.3-codex-xhigh Model | N/A | N/A | N/A | N/A |\n| 75 | gpt-5.4-2026-03-05-medium Model | N/A | N/A | N/A | N/A |\n| 76 | gpt-5.5-2026-04-23-medium Model | N/A | N/A | N/A | N/A |\n| 77 | gpt-5.5-2026-04-23-xhigh Model | N/A | N/A | N/A | N/A |\n| 78 | gpt-oss-120b Model | N/A | N/A | N/A | N/A |\n| 79 | gpt-oss-120b-high Model | N/A | N/A | N/A | N/A |\n| 80 | gpt-oss-20b Model | N/A | N/A | N/A | N/A |\n| 81 | Grok 4 Model | N/A | N/A | N/A | N/A |\n| 82 | Grok Code Fast 1 Model | N/A | N/A | N/A | N/A |\n| 83 | horizon-alpha Model | N/A | N/A | N/A | N/A |\n| 84 | horizon-beta Model | N/A | N/A | N/A | N/A |\n| 85 | Kimi K2 Model | N/A | N/A | N/A | N/A |\n| 86 | Kimi K2 Instruct 0905 Model | N/A | N/A | N/A | N/A |\n| 87 | Kimi K2 Thinking Model | N/A | N/A | N/A | N/A |\n| 88 | Kimi K2.5 Model | N/A | N/A | N/A | N/A |\n| 89 | Kimi K2.6 Model | N/A | N/A | N/A | N/A |\n| 90 | Llama-3.3-70B-Instruct Model | N/A | N/A | N/A | N/A |\n| 91 | Llama-4-Maverick-17B-128E-Instruct Model | N/A | N/A | N/A | N/A |\n| 92 | Llama-4-Scout-17B-16E-Instruct Model | N/A | N/A | N/A | N/A |\n| 93 | MiniMax M2 Model | N/A | N/A | N/A | N/A |\n| 94 | MiniMax M2.1 Model | N/A | N/A | N/A | N/A |\n| 95 | MiniMax M2.5 Model | N/A | N/A | N/A | N/A |\n| 96 | MiniMax M2.7 Model | N/A | N/A | N/A | N/A |\n| 97 | o3-2025-04-16 Model | N/A | N/A | N/A | N/A |\n| 98 | o4-mini-2025-04-16 Model | N/A | N/A | N/A | N/A |\n| 99 | Qwen2.5-72B-Instruct Model | N/A | N/A | N/A | N/A |\n| 100 | Qwen2.5-Coder-32B-Instruct Model | N/A | N/A | N/A | N/A |\n| 101 | Qwen3-235B-A22B Model | N/A | N/A | N/A | N/A |\n| 102 | Qwen3-235B-A22B no-thinking Model | N/A | N/A | N/A | N/A |\n| 103 | Qwen3-235B-A22B thinking Model | N/A | N/A | N/A | N/A |\n| 104 | Qwen3-235B-A22B-Instruct-2507 Model | N/A | N/A | N/A | N/A |\n| 105 | Qwen3-235B-A22B-Thinking-2507 Model | N/A | N/A | N/A | N/A |\n| 106 | Qwen3-30B-A3B-Instruct-2507 Model | N/A | N/A | N/A | N/A |\n| 107 | Qwen3-30B-A3B-Thinking-2507 Model | N/A | N/A | N/A | N/A |\n| 108 | Qwen3-32B Model | N/A | N/A | N/A | N/A |\n| 109 | Qwen3-32B no-thinking Model | N/A | N/A | N/A | N/A |\n| 110 | Qwen3-32B thinking Model | N/A | N/A | N/A | N/A |\n| 111 | Qwen3-Coder-30B-A3B-Instruct Model | N/A | N/A | N/A | N/A |\n| 112 | Qwen3-Coder-480B-A35B-Instruct Model | N/A | N/A | N/A | N/A |\n| 113 | Qwen3-Coder-Next Model | N/A | N/A | N/A | N/A |\n| 114 | Qwen3-Next-80B-A3B-Instruct Model | N/A | N/A | N/A | N/A |\n| 115 | Qwen3.5-27B Model | N/A | N/A | N/A | N/A |\n| 116 | Qwen3.5-397B-A17B Model | N/A | N/A | N/A | N/A |\n| 117 | Step-3.5-Flash Model | N/A | N/A | N/A | N/A |\n\n`Inspect`\n\nbutton.", "url": "https://wpnews.pro/news/13-models-and-4-agents-on-swe-tasks-go-java-python-rust-ts", "canonical_source": "https://swe-rebench.com", "published_at": "2026-07-31 15:28:28+00:00", "updated_at": "2026-07-31 15:52:51.565765+00:00", "lang": "en", "topics": ["artificial-intelligence", "machine-learning", "large-language-models", "ai-agents", "ai-research"], "entities": ["Fable 5", "Grok 4.5", "Opus 5", "GLM-5.2", "GPT-5.6 Sol", "Junie Agent", "Claude Code Agent", "Codex Agent"], "alternates": {"html": "https://wpnews.pro/news/13-models-and-4-agents-on-swe-tasks-go-java-python-rust-ts", "markdown": "https://wpnews.pro/news/13-models-and-4-agents-on-swe-tasks-go-java-python-rust-ts.md", "text": "https://wpnews.pro/news/13-models-and-4-agents-on-swe-tasks-go-java-python-rust-ts.txt", "jsonld": "https://wpnews.pro/news/13-models-and-4-agents-on-swe-tasks-go-java-python-rust-ts.jsonld"}}