← Models
gpt-oss-120b
evaluatedVendor: OpenAIQuant: MXFP4Type: local
Capability starfish
green = capability (pass-rate) · blue = operational (vs roster) · dashed = roster average
Overall
mean domain rate
Solved
45/48
latest per bench
Gauntlet time
1m
~2s/bench
Speed
47.1
tok/s
Steps/bench
1.4
decisiveness
$ / run
—
OpenRouter-equiv
Overall rank
#9 of 26Agentic
#19 of 22Coding
#11 of 22Speed (tok/s)
#5 of 22Decisive
#2 of 22Strongest
Gate · 100%Weakest
Agentic · 78%Where it fits · live, recomputed from current results
#9 of 26 overall; strongest on Gate (100%); weakest on Agentic (78%); #5/22 for speed (47.1 tok/s). (auto-generated from current results)
Verdict history · snapshotted daily when it changes
2026-07-09#9 of 26 overall; strongest on Gate (100%); weakest on Agentic (78%); #5/22 for speed (47.1 tok/s). (auto-generated from current results)
2026-07-08#10 of 26 overall; strongest on Gate (100%); weakest on Agentic (78%); #5/22 for speed (47.1 tok/s). (auto-generated from current results)
2026-07-07#16 of 28 overall; strongest on Gate (100%); weakest on SupportFlow (11%); #20/25 on the supportflow gauntlet; #6/26 for speed (45.5 tok/s). (auto-generated from current results)
2026-07-06#6 of 28 overall; strongest on Gate (100%); weakest on Agentic (78%); #6/26 for speed (47.1 tok/s). (auto-generated from current results)
2026-07-03#5 of 28 overall; strongest on Gate (100%); weakest on Agentic (78%); #6/26 for speed (47.1 tok/s). (auto-generated from current results)
2026-07-02#3 of 30 overall; strongest on Gate (100%); weakest on Agentic (78%); #5/29 for speed (47.1 tok/s). (auto-generated from current results)
2026-06-29#3 of 28 overall; strongest on Gate (100%); weakest on Agentic (78%); #5/27 for speed (47.1 tok/s). (auto-generated from current results)
Domain breakdown
Gate
100%
13/13
Coding
83%
5/6
Agentic
78%
7/9
C#
100%
15/15
PHP
100%
5/5
Benches run
green = passed · red = failed · hover for harness & time
Gate (13)
Coding (6)
Agentic (9)
C# (15)
PHP (5)
History (monthly pass-rate)
2026-06
91%
135/149
Availability record
When this model appeared in our catalogue census.
2026-06-25 · openrouter2026-06-25 · openrouter