gemini-2.5-flash-preview-05-20: benchmark scores, price and ranking
gemini-2.5-flash-preview-05-20 from Google DeepMind ranks #186 of 743 LLMs on the LLMs Tiger aggregate benchmark score (50.5), based on 5 benchmark results captured up to 2026-10-13.
- Organisation
- Google DeepMind
- Type
- Preview release
- Open weights
- No
- Released
- 2025-05-20 (source: name)
- Aggregate rank
- #186 of 743 (score 50.5)
Relative to other models, gemini-2.5-flash-preview-05-20 is strongest on Fiction.liveBench (#7 of 47) and weakest on EnigmaEval (#13 of 24).
Across the 5 benchmarks where at least ten models are scored, gemini-2.5-flash-preview-05-20 is in the top 10% on 0 and in the top half on 4.
gemini-2.5-flash-preview-05-20 benchmark results
| Benchmark | Score | Rank | Source | Captured |
|---|---|---|---|---|
| Fiction.liveBench | 77.8% | #7 of 47 | Epoch AI Benchmarking Hub | 2026-10-13 |
| Aider Polyglot | 55.1% | #16 of 49 | Aider polyglot leaderboard | 2026-10-13 |
| OTIS Mock AIME | 70.8% | #78 of 189 | Epoch AI Benchmarking Hub | 2026-10-13 |
| Humanity's Last Exam | 11.0% | #13 of 27 | Epoch AI Benchmarking Hub | 2026-10-13 |
| EnigmaEval | 2.7% | #13 of 24 | Epoch AI Benchmarking Hub | 2026-10-13 |
About gemini-2.5-flash-preview-05-20
gemini-2.5-flash-preview-05-20 is a preview release from Google DeepMind. It was released on 2025-05-20, according to the model's own name. No API price is listed for it yet. We hold 5 benchmark results for it across 3 areas, and none is in the top tenth of its board.
Where gemini-2.5-flash-preview-05-20 stands
On our aggregate score, gemini-2.5-flash-preview-05-20 is #186 of 743 ranked models, around the middle of the table. Its score is 50.5; the scale is centred on 50, so it sits 0.5 points above the typical model. That beats 75% of the ranked field. Among the 41 ranked Google DeepMind models it is #11. The next one up is Gemini 3 Pro at 50.6. The next one down is gemini-2.5-pro_16K at 50.4.
gemini-2.5-flash-preview-05-20, benchmark by benchmark
Fiction.liveBench puts gemini-2.5-flash-preview-05-20 at 77.8%, #7 of 47. That beats 85% of the models tested, but 16.6 points short of grok-4-0709 (94.4%). The results either side are QwQ-32B (83.3%) and DeepSeek-R1 (May 2025) (75.0%). It covers long-context story comprehension (16k). The board median is 61.1%, so gemini-2.5-flash-preview-05-20 is 16.7 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
On Aider Polyglot, gemini-2.5-flash-preview-05-20 reaches 55.1%. It ranks #16 of 49, ahead of 67% of the field, and 28.0 points behind gemini-2.5-pro-preview-06-05 at 83.1%. The results either side are DeepSeek V3 0324 (55.1%) and Quasar Alpha (54.7%). It covers edits across 6 languages (225 exercises). The board median is 44.9%, so gemini-2.5-flash-preview-05-20 is 10.2 points above it. Source: Aider polyglot leaderboard, captured 2026-10-13.
gemini-2.5-flash-preview-05-20 manages 70.8% on OTIS Mock AIME, good for #78 of 189. Claude Fable 5.1 (max) leads at 100.0%, 29.2 points further up, and 59% of models sit below. The results either side are MiniMax-M3 (71.1%) and qwen3-30b-a3b-thinking-2507 (70.3%). It covers competition maths (AIME 2024-25 style). The board median is 57.8%, so gemini-2.5-flash-preview-05-20 is 13.1 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
Humanity's Last Exam puts gemini-2.5-flash-preview-05-20 at 11.0%, #13 of 27. That beats 52% of the models tested, but 29.6 points short of Muse Spark (40.6%). The results either side are Claude Opus 4.1 (unknown thinking) (11.5%) and Claude Opus 4 (10.7%). It covers expert-level questions across domains. The board median is 10.7%, so gemini-2.5-flash-preview-05-20 is 0.2 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
gemini-2.5-flash-preview-05-20 lands below the middle on EnigmaEval: 2.7%, #13 of 24, with 50% of the field scoring higher. Gemini 3 Pro Preview leads at 18.2%; the closest result above is Claude Sonnet 4 (unknown thinking) at 3.1%. It covers multi-step puzzle hunts. The board median is 2.9%, so gemini-2.5-flash-preview-05-20 is 0.2 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
Strengths and weaknesses
Reasoning, maths and science: solid, ahead of 52% of the field on average across 3 benchmarks. Its best board here is OTIS Mock AIME (#78 of 189) and its weakest is EnigmaEval (#13 of 24).
Release history
The Google DeepMind model released before it was Gemini 2.5 Pro Preview (Jun 2025) on 2025-05-06, which scores 51.0. The next one was gemini-2.5-flash-preview-05-20_16K on 2025-05-20, at 49.3.
Reading these numbers
The 5 scores on this page come from 2 sources: Epoch AI Benchmarking Hub (4), Aider polyglot leaderboard (1). 1 of them come from an official leaderboard and 4 from datasets or third-party runs. With 5 results the aggregate rank can move a lot when one new score arrives.
Models ranked nearby
- #183 GPT-5-codex (high) (50.5)
- #184 grok-3-mini-beta (high) (50.5)
- #185 DeepSeek Chat V3 (prev) (50.5)
- #187 Kimi K2 Thinking Turbo (50.5)
- #188 swe-search (50.5)
- #189 Qwen3.7 Flash (50.5)
More from Google DeepMind
- Gemini 3.8 Flash (high) (#26, 54.0)
- Gemini 3.7 Flash (high) (#27, 53.9)
- Gemini 3 Pro Preview (#31, 53.7)
- Gemini 3 Deep Think (#60, 52.6)
- AI co-mathematician (#69, 52.3)
- Gemini 3 Flash (#120, 51.5)
Compare it on the live leaderboard → · How the aggregate score works