BridgeBenchBridgeBench

GPT-OSS 120B

Rank #2 · 120B · FP8

Summary

Pass Rate

66.7%

Tasks Passed

90/135

Model Size

120B

Quantization

FP8

Median Throughput

41.9 tok/s

Median TTFT

498 ms

Inference Success

100.0%

Avg Latency

11550 ms

Hardware Profile

Device

DGX Spark

Chip

GB10 Grace Blackwell

Memory

128 GB Unified

Backend

ollama

Quantization

FP8

Peak GPU Mem

0.0 GB

Category Results

Speed
41.9 tok/s · 498ms TTFT
Hallucination
10/30
33.3%
Code Generation
14/20
70.0%
Reasoning
26/30
86.7%
Instruction Following
16/20
80.0%

Task Results

Speed20/20 passed
TaskDifficultyResultLatencyTokens
ttft-short-100standardPass2900ms39
ttft-short-200standardPass2757ms100
ttft-medium-500standardPass2710ms100
ttft-medium-1kstandardPass2860ms100
ttft-long-2kstandardPass2960ms100
ttft-chat-contextstandardPass2702ms100
ttft-json-outputstandardPass2781ms100
ttft-multilangstandardPass2776ms100
ttft-reasoningstandardPass2784ms100
ttft-creativestandardPass2706ms100
tp-essaystandardPass48156ms2000
tp-code-appstandardPass48200ms2000
tp-tutorialstandardPass36231ms1500
tp-analysisstandardPass36314ms1500
tp-debugstandardPass36094ms1500
tp-architecturestandardPass87910ms2000
tp-comparisonstandardPass48214ms2000
tp-securitystandardPass48146ms2000
tp-algorithmstandardPass48293ms2000
tp-documentationstandardPass48185ms2000
Hallucination10/30 passed
TaskDifficultyResultLatencyTokens
fact-01easyPass14148ms93
fact-02mediumPass3073ms114
fact-03mediumFail13802ms487
fact-04hardFail12168ms500
fact-05hardFail6450ms261
fact-06easyFail5749ms153
fact-07mediumPass4176ms91
fact-08hardPass6853ms276
fact-09mediumPass2930ms111
fact-10hardPass12170ms500
code-01easyFail8561ms345
code-02mediumPass4035ms154
code-03mediumPass12212ms500
code-04hardFail12242ms500
code-05hardFail12220ms500
code-06easyFail12147ms500
code-07mediumFail12169ms500
code-08hardFail12185ms500
code-09mediumFail7418ms298
code-10hardPass12242ms500
cal-01mediumFail12242ms500
cal-02hardFail12187ms500
cal-03mediumFail12181ms500
cal-04easyFail12164ms500
cal-05hardFail12241ms500
cal-06mediumFail13947ms500
cal-07hardFail12236ms500
cal-08mediumFail12251ms500
cal-09easyPass12183ms500
cal-10hardFail12255ms500
Code Generation14/20 passed
TaskDifficultyResultLatencyTokens
fn-01easyPass3316ms124
fn-02easyPass4398ms172
fn-03mediumFail2799ms103
fn-04mediumPass5941ms236
fn-05mediumPass18241ms755
fn-06hardFail10799ms438
fn-07hardFail8232ms333
fn-08hardFail9370ms377
bug-01easyPass4174ms154
bug-02mediumPass4341ms155
bug-03hardPass7321ms286
bug-04mediumPass5483ms210
algo-01mediumPass5495ms216
algo-02hardPass6167ms245
algo-03mediumPass3489ms131
algo-04hardPass4766ms185
multi-01hardPass6912ms275
multi-02hardFail11538ms468
multi-03hardPass9331ms377
multi-04hardFail8087ms325
Reasoning26/30 passed
TaskDifficultyResultLatencyTokens
arith-01hardPass4461ms173
arith-02hardPass2011ms71
arith-03expertPass5310ms207
arith-04expertPass7221ms289
arith-05expertPass2182ms76
arith-06hardPass8832ms357
spatial-01hardPass7163ms285
spatial-02expertPass4194ms159
spatial-03expertPass15730ms646
spatial-04hardPass10807ms438
spatial-05expertPass4721ms181
spatial-06hardFail24251ms1000
cstr-01hardPass5154ms200
cstr-02expertPass14177ms577
cstr-03expertFail24363ms1000
cstr-04hardPass24247ms1000
cstr-05expertFail78398ms1000
cstr-06hardPass18216ms743
adv-01hardPass4081ms156
adv-02expertPass4337ms167
adv-03expertPass7043ms280
adv-04hardPass11064ms371
adv-05expertPass5857ms233
adv-06expertPass7819ms315
cf-01hardPass5658ms222
cf-02expertFail9732ms390
cf-03expertPass7095ms283
cf-04hardPass10323ms417
cf-05expertPass17652ms724
cf-06expertPass11225ms454
Instruction Following16/20 passed
TaskDifficultyResultLatencyTokens
fmt-01easyPass1962ms67
fmt-02easyPass3964ms154
fmt-03mediumPass3487ms135
fmt-04mediumPass2791ms103
fmt-05hardPass3911ms150
fmt-06hardPass6781ms272
con-01easyPass11054ms453
con-02easyFail2891ms108
con-03mediumPass3605ms69
con-04mediumPass2805ms106
con-05hardPass8812ms357
con-06hardFail12325ms500
role-01mediumPass4397ms165
role-02mediumPass12198ms500
role-03hardPass12329ms500
role-04hardFail12225ms500
mc-01hardFail12237ms500
mc-02hardPass8329ms331
mc-03hardPass12128ms494
mc-04hardPass10471ms424