← Claude Sonnet 5 · API · High effort (default)
Evidenceacceptedthird_party_leaderboard
SWE-bench Verified
Vals.ai SWE-bench Verified (mini-swe-agent, bash-only)
ev_2026_07_sonnet5_swe_vals
Result
79.6%
Confidence A
- Tested
- 2026-07-09
- Runs
- 1
Configuration
How it was measured
- Exact model
- claude-sonnet-5
- Reasoning
- high
- Agent harness
- mini-swe-agent
- Tool access
- bash only (no specialized editor tools)
- Prompting
- Vals default provider config; max tokens raised to model max
Independent third-party SWE-bench Verified board (Vals.ai) using mini-swe-agent for all models. Prefer this over vendor SWE-bench Pro for comparative STR.
Caveats
Known limitations
- Single public board run; variance not fully published as multi-trial CI on this scrape.
- High/max effort config — not bare non-reasoning chat.
- mini-swe-agent is intentionally minimal; richer harnesses can score higher.
- Do not compare this 79.6% Verified score directly to vendor SWE-bench Pro 63.2%.
Weighting
Evidence weight factors
The engine weights each evidence row by these factors (0–1) when fusing scores — higher is more trusted.
relevance0.95
quality0.85
recency1.00
health0.80
comparability0.90
reliability0.85
Benchmark
SWE-bench Verified
Human-filtered 500-instance subset of SWE-bench evaluating whether models can resolve real GitHub issues in existing repositories under a controlled agent harness.
Benchmark source →