LLM Reference
activeCoding

SWE-bench Pro

Metric: % Resolved (higher is better)Introduced: 2025

731-task multilingual real-world GitHub issue benchmark extending SWE-bench Verified with harder, more diverse tasks across Python, JavaScript, TypeScript, Java, Go, C++, and Rust. High benchmark score alone doesn't make a model the right pick — weigh it against pricing, API availability, and release date.

Models ranked

42

tracked on this benchmark

Score band

80.3 – 38.7

best → lowest tracked

Snapshot trend

+14.50

Jul 8 → Jul 24 · 1 models

Leaderboard

Tracked models ranked by % Resolved (higher is better).

Compare candidates
#Model variant and provenanceScore
1
Claude Fable 5
Version: SWE-bench ProHarness: Not recordedEvaluator: Not recordedObserved: Jun 9, 2026Confidence: Not recordedSource

Notes: DAT-5842: Unstarred in Anthropic's shared Fable 5/Mythos 5 launch table, so this is a Claude Fable 5 score.

80.3
2
Claude Opus 5

Configuration: Claude Opus 5 standard configuration

Version: SWE-bench ProHarness: Anthropic system-card standard configuration; average of five trialsEvaluator: AnthropicObserved: Jul 24, 2026Confidence: confirmedSource

Notes: Vendor-reported. Source: p.149, Table 8.1.A. This is distinct from SWE-bench Verified.

79.2
3
Claude Opus 4.8
Version: SWE-bench ProHarness: Not recordedEvaluator: Not recordedObserved: May 28, 2026Confidence: Not recordedSource

Notes: Official Anthropic launch benchmark.

69.2
4
Grok 4.5
Version: SWE Bench Pro resolve rate (xAI first-party reported chart)Harness: Not recordedEvaluator: Not recordedObserved: Jul 8, 2026Confidence: Not recordedSource

Notes: xAI launch blog SWE Bench Pro chart reports Grok 4.5 at 64.7% resolve rate. Evaluator/source: xAI first-party launch chart only; blog does not specify Pass@1 vs other metric, agent scaffold, reasoning level, or public SWE-bench Pro harness for Grok 4.5 (separate token-efficiency section references SWE Bench Pro tasks but not this score's run config). Variant: Grok 4.5 (harness unspecified). Harness status: not specified in primary source — treat as directional vendor-reported chart, not reproducible from seed metadata. Confidence: medium for reported value, low for reproducibility/comparability. Recommended seed value: 64.7 pending harness documentation or independent leaderboard confirmation.

64.7
5
GPT-5.6 Sol
Version: SWE-Bench Pro; OpenAI GA launch table; codingHarness: Not recordedEvaluator: Not recordedObserved: Jul 9, 2026Confidence: Not recordedSource

Notes: Official GPT-5.6 GA launch benchmark row.

64.6
6
Claude Opus 4.7
Version: SWE-bench Pro (pass@1)Harness: Not recordedEvaluator: Not recordedObserved: Jun 7, 2026Confidence: Not recordedSource
64.3
7
Claude Sonnet 5
Version: SWE-bench Pro; Anthropic standard configuration, adaptive thinking at max effort, default sampling, average over 5 trialsHarness: Not recordedEvaluator: Not recordedObserved: Jun 30, 2026Confidence: Not recordedSource
63.2
8
GLM-5.2
Version: SWE-bench Pro (% resolved, self-reported)Harness: Not recordedEvaluator: Not recordedObserved: Jun 13, 2026Confidence: Not recordedSource
62.1
9
Qwen3.7-Max
Version: Not recordedHarness: Not recordedEvaluator: Not recordedObserved: May 20, 2026Confidence: Not recordedSource
60.6
10
LongCat-2.0
Version: SWE-bench Pro; Code Agent chart; LongCat launch benchmarkHarness: Not recordedEvaluator: Not recordedObserved: Jun 30, 2026Confidence: Not recordedSource

Notes: Official LongCat-2.0 launch chart reports 59.51. Evaluator/harness details are not fully reproducible from public text; keep separate from independent SWE-bench Verified. Confidence: confirmed for reported value, medium for reproducibility.

59.5
11
MiniMax M3
Version: MiniMax-reportedHarness: Not recordedEvaluator: Not recordedObserved: Jun 1, 2026Confidence: Not recordedSource
59.0
12
Nex-N2-Pro
Version: SWE-bench Pro — self-reportedHarness: Not recordedEvaluator: Not recordedObserved: Jun 4, 2026Confidence: Not recordedSource

Notes: DAT-5798: Self-reported in model card and SiliconFlow blog. Not independently verified.

58.8
13
GPT-5.5
Version: SWE-bench Pro (pass@1)Harness: Not recordedEvaluator: Not recordedObserved: Jun 7, 2026Confidence: Not recordedSource
58.6
14
Kimi K2.6
Version: SWE-bench Pro (resolved%)Harness: Not recordedEvaluator: Not recordedObserved: Jun 7, 2026Confidence: Not recordedSource
58.6
15
GPT-5.5 Pro
Version: SWE-bench Pro; standard GPT-5.5 weightsHarness: Not recordedEvaluator: Not recordedObserved: Apr 24, 2026Confidence: Not recordedSource

Notes: Medium confidence secondary-source row for GPT-5.5 standard; attributed to Pro because the Pro product uses the same underlying weights.

58.6
16
GLM-5.1
Version: Not recordedHarness: Not recordedEvaluator: Not recordedObserved: Apr 7, 2026Confidence: Not recordedSource
58.4
17
GPT-5.4
Version: SWE-bench Pro Public (pass@1)Harness: Not recordedEvaluator: Not recordedObserved: Jun 7, 2026Confidence: Not recordedSource
57.7
18
Xiaomi MiMo-V2.5-Pro
Version: SWE-bench Pro (resolved%)Harness: Not recordedEvaluator: Not recordedObserved: Jun 7, 2026Confidence: Not recordedSource
57.2
19
GPT-5.3-Codex
Version: SWE-bench Pro Public (pass@1)Harness: Not recordedEvaluator: Not recordedObserved: Jun 7, 2026Confidence: Not recordedSource
56.8
20
GPT-5.2 Codex
Version: SWE-Bench Pro Public; OpenAI launch comparison; GPT-5.2-Codex xhighHarness: Not recordedEvaluator: Not recordedObserved: Dec 18, 2025Confidence: Not recordedSource

Notes: OpenAI GPT-5.3-Codex appendix reports GPT-5.2-Codex 56.4%, GPT-5.2 55.6%, GPT-5.1-Codex-Max 50.8%. Evaluator: OpenAI self-reported launch benchmark. Harness details are not fully reproducible from public text; keep distinct from independent SWE-bench Verified. Confidence: confirmed for reported value, medium for reproducibility.

56.4
21
Step 3.7 Flash
Version: 2nd at release (Claude Opus 4.7: 64.3, GPT-5.5: 58.6)Harness: Not recordedEvaluator: Not recordedObserved: May 29, 2026Confidence: Not recordedSource
56.3
22
MiniMax M2.7 Highspeed
Version: SWE-Pro (resolved%)Harness: Not recordedEvaluator: Not recordedObserved: Jun 7, 2026Confidence: Not recordedSource
56.2
23
MiniMax M2.7
Version: OpenRouter SWE-Pro listingHarness: Not recordedEvaluator: Not recordedObserved: Jun 1, 2026Confidence: Not recordedSource
56.2
24
GPT-5.2
Version: Not recordedHarness: Not recordedEvaluator: Not recordedObserved: Dec 9, 2025Confidence: Not recordedSource
55.6
25
DeepSeek V4 Pro
Version: Not recordedHarness: Not recordedEvaluator: Not recordedObserved: Apr 24, 2026Confidence: Not recordedSource
55.4
Showing the top 25 of 42 tracked models. Browse all models.

How to read this benchmark

This benchmark scores models where higher is better. Scores are useful for directional filtering and shortlisting — not for universal quality ranking. Prefer benchmarks closest to your workload, then validate the linked model pages for pricing, context window, and provider availability.

Trust this score when

  • There is a fresh timestamped snapshot (or multiple snapshots) for this benchmark.
  • The model list covers the same version family you can actually deploy today.
  • Top candidates overlap with your required routing and feature requirements.

Be cautious when

  • There is only one benchmark snapshot or the dataset appears stale.
  • The benchmark metric direction is opposite of your decision objective.
  • The score difference between options is narrow and likely within implementation variance.

FAQ

What does the benchmark measure?

731-task multilingual real-world GitHub issue benchmark extending SWE-bench Verified with harder, more diverse tasks across Python, JavaScript, TypeScript, Java, Go, C++, and Rust. On this page it lists 42 tracked model variants where higher is better.

Is a higher SWE-bench Pro score always better?

For this benchmark, higher is better. A high score helps you shortlist, but confirm pricing, context window, and provider availability on each model page before committing — the top scorer is not always the right pick for your workload or budget.

How current is this SWE-bench Pro data?

This benchmark was last reviewed on Apr 15, 2026. The tracked score average moved +14.50 points across the last 3 snapshots.

Related benchmarks

Last reviewed: Apr 15, 2026