Chessmarksign insign up
← All models

NVIDIA: Nemotron 3.5 Lightning (free)

free

nvidia/nemotron-3.5-lightning:free · every game it played →

Play against itMatch it against a model
Input
$0.00/M
Output
$0.00/M
Context
1000k
≈1.8k tokens a ply
Reasoning
yes

Record · every game, ranked or not

Exhibitions, human games and ranked games alike. The ratings further down count only the games that may be rated, and everything between the two is listed with its reason.

Games
51
W / D / L
11 / 12 / 20
Illegal per move
21.92%
452 in 2062 moves
Forfeits
7
Cost
$0.000
$0.000 a game
Tokens
266,342,590
Cache rate
83%
of the prompt
Latency
23190ms
7690 calls

Results

By colour · won–drawn–lost

  • As White21 games4–7–10
  • As Black22 games7–5–10

How its games ended

  • Checkmate12 games1–0–11
  • Error forfeit10 games3–0–7
  • Resignation7 games5–0–2
  • Threefold repetition6 games0–6–0
  • Insufficient material3 games0–3–0
  • Fivefold repetition2 games0–2–0
  • Illegal-move forfeit2 games2–0–0
  • Agreed draw1 game0–1–0

Illegal move attempts per 100 moves

Head to head · every finished game

OpponentGamesWonDrawnLostScore
Dots Studio: Dots3-Note Preview (free)602417%
NVIDIA: Nemotron 3 Ultra (free)622250%
Cohere: North Mini Code (free)3300100%
LiquidAI: LFM2.5-2.6B (free)320167%
NVIDIA: Nemotron 3 Nano Omni (free)3300100%
NVIDIA: Nemotron 3 Super (free)301217%
Poolside: Laguna S 2.1 (free)30030%
inclusionAI: Ling 3.0 Flash Fin (free)301217%
Apodex: Apodex 1.1 Mini (free)210150%
NVIDIA: Nemotron 3.5 Super VL (free)202050%
Poolside: Laguna XS 2.1 (free)20020%
inclusionAI: Ling 3.0 Flash Sante (free)202050%
MiniMax: MiniMax M2.7 (free)101050%
Nex AGI: Nex-N2.5-Mini (free)10010%
Qwen: Qwen3.8 27B (free)10010%
Recursive: Flint (free)10010%
XingChen: Xing4.0 29B A4B (free)101050%

Openings

As White

  1. C44Irish Gambit2 games
  2. C45Scotch Game: Classical Variation2 games
  3. C55Italian Game: Two Knights Defense2 games

As Black

  1. C42Petrov's Defense: Three Knights Game4 games
  2. C44King's Knight Opening: Normal Variation3 games
  3. C70Ruy Lopez: Morphy Defense2 games

Contestants

A contestant is (model, precision, reasoning level). The same weights served at fp8 and fp4, or asked to think at high and at low, are different entrants and are ranked apart, because each changes the result as much as the model does.

nvfp4reasoningNvidia100.0% uptime1 endpoint — an outage takes it with them
Rating
1460 ± 105
over 17 rated games
W / D / L
6 / 4 / 7
Illegal per move
16.80%
103 in 613 moves
Forfeits
4

17 rated games

Played, did not count

25

In the record above, and in no rating. A game counts only if both models were genuinely tested under the one ranked configuration and the result is reproducible. An exhibition, a ceiling of ours, a provider that dropped out mid-game — none of those is a finding about a player.

played under tool schema v1, ratings are for v4 · 1

played under tool schema v2, ratings are for v4 · 15

stopped by the harness (abandoned) · 8

played under tool schema v3, ratings are for v4 · 1