Home Model series Gemini Flash Gemini Flash · DeepSWE · v1.1 · Release history Gemini Flash · DeepSWE · v1.1 · Release history View scores, configurations, and disclosures for DeepSWE · v1.1 across 3 Gemini Flash releases, ordered by release date.
Differences are calculated only for unique, active, exact values between direct predecessors when the unit, metric, and explicitly recorded configuration and conditions match.
2026-07-21 · 1 score record for this release, including 1 exact value (may include historical records)
Gemini 3.6 Flash · DeepSWE scores and configurations Metric and score Configuration and evaluation conditions Sources Score 49% Reported 49% · Unit % Published Unconfirmed
Gemini 3.6 Flash high thinking Settings Gemini API gemini-3.6-flash; high thinking for DeepSWE; default sampling elsewhere. Unknowns Exact thinking budget is not published. Tools Tool access not recorded Reasoning budget high Framework not recorded Protocol DeepSWE v1.1 public leaderboard; highest reported Gemini thinking level is high. Conditions Main leaderboard; pass@1. View this score in rankings
PDF p.2, DeepSWE v1.1 row Attachment · Gemini 3.6 Flash evaluation methodology and results 2026-08-13 · 1 score record for this release, including 1 exact value (may include historical records)
Records for the direct predecessor do not provide matching conditions for a valid difference.
Gemini 3.7 Flash · DeepSWE scores and configurations Metric and score Configuration and evaluation conditions Sources Score 65.3% Reported 65.3% · Unit % Published Unconfirmed
Gemini 3.7 Flash high thinking Settings Gemini API gemini-3.7-flash; high thinking on DeepSWE; default sampling. Unknowns Exact thinking budget is not published. Tools Tool access not recorded Reasoning budget high Framework mini-swe-agent Protocol Self-computed with mini-swe-agent and LiteLLM 1.96; high thinking. Conditions Gemini API; pass@1; results as of August 2026. View this score in rankings
PDF p.4, DeepSWE v1.1 row Attachment · Gemini 3.7 Flash evaluation methodology and results 2026-09-02 · 1 score record for this release, including 1 exact value (may include historical records)
Records for the direct predecessor do not provide matching conditions for a valid difference.
Gemini 3.8 Flash · DeepSWE scores and configurations Metric and score Configuration and evaluation conditions Sources Score 73.7% Reported 73.7% · Unit % Published Unconfirmed
Gemini 3.8 Flash high thinking Settings Gemini API gemini-3.8-flash; high thinking for DeepSWE; pass@1. Unknowns Exact thinking budget is not published. Tools Tool access not recorded Reasoning budget high Framework mini-swe-agent Protocol Self-computed with mini-swe-agent; highest reported thinking level, high; pass@1. Conditions Gemini API; results as of September 2026. View this score in rankings
PDF p.4, DeepSWE v1.1 row, Gemini 3.8 Flash column Attachment · Gemini 3.8 Flash evaluation methodology and results