Gemini 3.1 Pro Preview
Sold by 6 sellers, timed at 90.4 tok/s, scored 68.8 on the Coding Index, on 25 Design Arena boards, and with results on 2 openly licensed boards.
retrieved sha256 · 0bb53a77713f
The same model sold by 2 providers, each at its own list price. OpenRouter's rate is its routed rate across whichever of them it uses, and it need not match any one of them.
What sellers charge
6 sellers · 3.6x from cheapest to dearestOne tick per seller, on a log scale of input price per million from $1.00 at Google to $3.60 at Google AI Studio. 2 sellers undercut the $2.00 listed; 2 are dearer. A faint tick declares no number format. Every seller is named on the endpoint ledger below.
Measured speed
20,704 requests · 5 endpoints · 30-minute windowOpenRouter's own timing of its routed traffic, weighted by requests, over one 30-minute window read . Every route is on the endpoint ledger below.
Every endpoint
6 priced · 5 timed · cheapest first- Googlegoogle-vertex/global/flex$1.00$6.00$0.10not declared66knot matchednonenonenot matched11.11%
- Google AI Studiogoogle-ai-studio/flex$1.00$6.00$0.10not declared66knot matchednonenonenot matched100%
- Googlegoogle-vertex/global · listed$2.00$12.00$0.20not declared66knot matchednonenonenot matched100%
- Google AI Studiogoogle-ai-studio · listed$2.00$12.00$0.20not declared66knot matchednonenonenot matched99.77%
- Googlegoogle-vertex/global/priority$3.60$21.60$0.36not declared66knot matchednonenonenot matched100%
- Google AI Studiogoogle-ai-studio/priority$3.60$21.60$0.36not declared66knot matchednonenonenot matchedno reading
- Googlenot matchednot matchednot matchednot matchedno offer88 tok/s452.85s19,026 reqno offer
- Googlenot matchednot matchednot matchednot matchedno offer80.5 tok/s532.15s2 reqno offer
- Google AI Studionot matchednot matchednot matchednot matchedno offer117 tok/s324.74s1,529 reqno offer
- Google AI Studionot matchednot matchednot matchednot matchedno offer120 tok/s90.43.68s146 reqno offer
- Google AI Studionot matchednot matchednot matchednot matchedno offer161 tok/s15510s1 reqno offer
Each track runs from a route's p90, the rate nine in ten of its requests met or beat, up to its median. Weighted by requests, the figure above reports where the traffic went rather than what one route does: 2 of the 5 timed routes are slower than it, and the middle route runs at 117 tok/s. The widest median-to-p90 gap is Google AI Studio's, 3.66x.
Google publishes 3 offers and had 2 routes timed; Google AI Studio publishes 3 offers and had 3 routes timed. The two documents share no route id, so those offers and routes are printed apart rather than paired.
Uptime is OpenRouter's own figure for each seller over the trailing day, read once a day, and the column prints the latest reading, taken . 5 of the 6 sellers with a reading fell under 99% on at least one of the 23 days held, 23 Oct 2026 to 21 Nov 2026: Google (google-vertex/global/flex) at 0.00% on 25 Oct 2026, Google AI Studio (google-ai-studio) at 91.79% on 26 Oct 2026, Google AI Studio (google-ai-studio/flex) at 97.09% on 10 Nov 2026, Google (google-vertex/global/priority) at 98.07% on 3 Nov 2026, and 1 more. Source: OpenRouter's public endpoints API, one reading a day for each seller: 1,211 sellers carried one on the latest of the 23 days held.
Source: OpenRouter's public endpoints API. 1257 sellers across 359 models when the mirror ran, 186 of them served by more than one. List prices, not what anybody paid; a discount is the seller's own. Read . Speed is OpenRouter's own timing over one 30-minute window, read .
What each seller accepts
12 parameters · 6 sellers- include_reasoningevery seller
- max_tokensevery seller
- reasoningevery seller
- reasoning_effortevery seller
- response_formatevery seller
- seedevery seller
- structured_outputsevery seller
- temperatureevery seller
- tool_choiceevery seller
- toolsevery seller
- top_pevery seller
- stop3 of 6not Google AI Studio (google-ai-studio), Google AI Studio (google-ai-studio/flex), and Google AI Studio (google-ai-studio/priority)
11 of these 12 are accepted by every one of the 6 sellers that publish a list; the other 1 are declared by some sellers and not others, so what a request sending one of them gets depends on which seller serves it. Each list is the seller's own declaration in OpenRouter's endpoints API, read , and nothing here tested one.
Price position
342 priced modelsBlended $4.50 per million tokens, 3 in to 1 out: cheaper than 42 of the 342 priced models in the catalog. Each bar is one stretch of the catalog on a logarithmic price scale; the mark is this listing.
Measured quality
145 models with an index · 132 on a design arena boardEach mark is one board this model stands on, placed at the share of that board it scores above, and led to its number in the list under the rule. The columns and the ledger below carry the figures.
- Agentic game development · Agents1060 Elowin rate 43.9% · source rank 23the board, 25 ranked
- Agentic HTML slides · Agents1226 Elowin rate 55.8% · source rank 4the board, 9 ranked
- Agentic slides · Agents1112 Elowin rate 33.8% · source rank 7the board, 9 ranked
- Agentic slides, HTML · Agents1219 Elowin rate 54.4% · source rank 4the board, 9 ranked
- Agentic slides, python-pptx · Agents1107 Elowin rate 33.9% · source rank 7the board, 9 ranked
- Android native · Agents1035 Elowin rate 41.1% · source rank 40the board, 41 ranked
- blender3d · Agents1037 Elowin rate 39.8% · source rank 18the board, 25 ranked
- codetovideo · Agents1133 Elowin rate 44.3% · source rank 19the board, 26 ranked
- Full stack · Agents1041 Elowin rate 40.4% · source rank 35the board, 44 ranked
- Godot game development · Agents1131 Elowin rate 53.8% · source rank 24the board, 34 ranked
- HTML slides · Agents1050 Elowin rate 48.4% · source rank 18the board, 25 ranked
- Mobile apps · Agents1085 Elowin rate 42.2% · source rank 44the board, 49 ranked
- pdf · Agents1107 Elowin rate 38.2% · source rank 32the board, 36 ranked
- PPTX slides · Agents1110 Elowin rate 34.1% · source rank 7the board, 8 ranked
- python-pptx slides · Agents1109 Elowin rate 31.9% · source rank 25the board, 29 ranked
- spreadsheet · Agents918 Elowin rate 28.6% · source rank 33the board, 32 ranked
- Web apps · Agents1106 Elowin rate 44.1% · source rank 36the board, 47 ranked
- 3D scenes · Models1214 Elowin rate 58.5% · source rank 48the board, 113 ranked
- ASCII art · Models1253 Elowin rate 63.5% · source rank 16the board, 67 ranked
- Code · Models1232 Elowin rate 64.4% · source rank 49the board, 120 ranked
- Data visualisation · Models1235 Elowin rate 61.8% · source rank 40the board, 110 ranked
- Game development · Models1191 Elowin rate 52.4% · source rank 62the board, 120 ranked
- SVG · Models1274 Elowin rate 67.9% · source rank 15the board, 83 ranked
- UI components · Models1268 Elowin rate 69.4% · source rank 32the board, 116 ranked
- Websites · Models1243 Elowin rate 64.4% · source rank 47the board, 126 ranked
Epoch AI has not placed Gemini 3.1 Pro Preview on its Capabilities Index, or places it under a name the pairing does not reach. Epoch places a model once four of its benchmarks carry a result.
Benchmark results
2 of 6 boards · 41 results| Task | All runs | Effort | Score |
|---|---|---|---|
| Chess Puzzles | 167 | high | 0.51 |
| EBR-bench | 34 | unstated | 0.14 |
| FrontierMath-2025-02-28-Private | 101 | unstated | 0.37 |
| FrontierMath-2025-02-28-Public | 87 | unstated | 0.89 |
| FrontierMath-Tier-4-2025-07-01-Private | 72 | unstated | 0.17 |
| FrontierMath-Tier-4-2025-07-01-Public | 48 | unstated | 0 |
| FrontierMath-Tier-4-v2-Private | 84 | unstated | 0.27 |
| FrontierMath-Tiers-1-3-v2-Private | 128 | unstated | 0.60 |
| Furniture Assembly | 42 | high | 0.27 |
| GPQA diamond | 342 | high | 0.94 |
| unstated | 0.94 | ||
| MirrorCode | 48 | high | 0.09 |
| Mystery Game Puzzles | 149 | high | 0.34 |
| medium | 0.32 | ||
| low | 0.29 | ||
| OTIS Mock AIME 2024-2025 | 319 | high | 0.96 |
| unstated | 0.96 | ||
| SimpleQA Verified | 106 | high | 0.73 |
Mean accuracy, 0 to 1, on the task the row names, as Epoch AI runs it and averaged over repeated runs. All eighteen of their tasks, including the held-out FrontierMath tiers, named as they name them. The model id carries the effort setting the run used. Epoch AI Benchmarking Hub names these runs gemini-3.1-pro-preview_high, gemini-3.1-pro-preview_medium, gemini-3.1-pro-preview_low and gemini-3.1-pro-preview, and the effort beside each score says which run it is, unstated where the publisher wrote none. A run missing from a task's rows has no published result there, which is not a score of zero. All runs is how many runs the publisher scored on the task, every model's. Read from epoch.ai ↗, which dates its own data . Epoch AI Benchmarking Hub data, by Epoch AI, under CC BY 4.0.
| Task | Score |
|---|---|
| AMPS_Hard | 98 |
| code_completion | 78.26 |
| code_generation | 74.65 |
| connections | 100 |
| consecutive_events | 85.24 |
| integrals_with_game | 78 |
| javascript | 59.09 |
| logic_with_navigation | 72 |
| math_comp | 96.08 |
| olympiad | 92.10 |
| paraphrase | 79.87 |
| plot_unscrambling | 74.13 |
| python | 50 |
| simplify | 71.05 |
| spatial | 98 |
| story_generation | 77.25 |
| summarize | 88.23 |
| tablejoin | 52.35 |
| tablereformat | 98.04 |
| theory_of_mind | 80.77 |
| typescript | 23.33 |
| typos | 82 |
| zebra_puzzle | 85.25 |
Percent correct on the task the row names, higher is better. All 23 published columns; there is no average to carry, because the site computes its headline in the browser from a weighting this does not reimplement. LiveBench names this run gemini-3.1-pro-preview-high. Read from livebench.ai ↗, which dates its own data . LiveBench, Apache-2.0. Results produced by the LiveBench maintainers on questions refreshed monthly.
2 of the 6 openly licensed boards Vitna mirrors carry a result for Gemini 3.1 Pro Preview, each printed as its publisher published it and none added to another. MTEB, HELM, Vectara hallucination leaderboard and Aider polyglot have not published one under a name that matches its listing, and names are matched to the catalog exactly and never guessed. A task that is a link opens its column on /capability, with every run scored on it; an effort that is a link opens its run's row there, where the run is ranked on Vitna's scale. Every board, with its licence and what its publisher measures, is on /benchmarks.
Source: OpenRouter (openrouter.ai/models), as of 17 Nov 16:29 UTC. sha256 0bb53a77713f over 853,270 bytes.
Catalog: OpenRouter's public model list.
Intelligence, coding and agentic indexes: Artificial Analysis, as published by OpenRouter.
Elo, win rate and rank: Design Arena, as published by OpenRouter.
Vitna ran none of these evaluations and adjusted none of these figures.
