source
BridgeBench
Benchmarks · /benchmarks/bridgebench-ttft
Speed TTFT
BridgeBench time-to-first-token leaderboard for coding-oriented workloads.
Passport
Verified but agingThis is an efficiency signal, so it belongs beside quality rather than being mistaken for quality.
metric
TTFT (ms)
judge
Speed / cost
direction
lower better
group id
bridgebench_ttft_2026_04
domain
Coding
What it measures vs what it misses
✓ Measures
Initial coding-response latency before tokens begin streaming.
✗ Misses
Final answer quality. Total task completion time.
Why this countsIt tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.Comparable-group ruleThis percentile only compares models inside the exact benchmark/version group shown here. It is not a universal score.What it missesIt does not fully capture repo-scale iteration, IDE ergonomics, or long debugging loops.
Leaderboard · this benchmark version
#1 · GPT-5.4 mini
BB · Undated
233.00ms
#2 · GPT-5.4
BB · Undated
397.00ms
#3 · Claude Opus 4.7
BB · Undated
852.00ms
#4 · GPT-5.5
BB · Undated
930.00ms
#5 · GPT-5.4 nano
BB · Undated
941.00ms
#6 · Claude Sonnet 4.6
BB · Undated
1207.00ms
#7 · Claude Opus 4.6
BB · Undated
1922.00ms
#8 · Grok 4.20
BB · Undated
1999.00ms
#9 · Grok 4.3
BB · Undated
3131.00ms
#10 · Grok 4
BB · Undated
3684.00ms
#11 · Gemini 3.1 Pro Preview
BB · Undated
7608.00ms